The 2nd Cross Company Respiratory Symposium (CCRS), held in Horsham, U.K. in 2012, brought together representatives from across the pharmaceutical industry with expert academics, in the common interest of improving the design and translational predictiveness of in vivo models of respiratory disease. Organized by the respiratory representatives of the European Federation of Pharmaceutical Industries and Federations (EFPIA) group of companies involved in the EU-funded project (U-BIOPRED), the aim of the symposium was to identify state-of-the-art improvements in the utility and design of models of respiratory disease, with a view to improving their translational potential and reducing wasteful animal usage. The respiratory research and development community is responding to the challenge of improving translation in several ways: greater collaboration and open sharing of data, careful selection of the species, complexity and chronicity of the models, improved practices in preclinical research, continued refinement in models of respiratory diseases and their sub-types, greater understanding of the biology underlying human respiratory diseases and their sub-types, and finally greater use of human (and especially disease-relevant) cells, tissues and explants. The present review highlights these initiatives, combining lessons from the symposium and papers published in Clinical Science arising from the symposium, with critiques of the models currently used in the settings of asthma, idiopathic pulmonary fibrosis and COPD. The ultimate hope is that this will contribute to a more rational, efficient and sustainable development of a range of new treatments for respiratory diseases that continue to cause substantial morbidity and mortality across the world.

INTRODUCTION

In vivo models provide investigators with the opportunity to assess the efficacy, pharmacodynamic, pharmacokinetic and toxicological profiles of novel interventions, in the complex milieu of the compromised organ of interest and the context of a whole body system. This warrants their inclusion in the pre-clinical critical pathways of many pre-clinical drug discovery programmes currently active in the respiratory setting. Furthermore, animal safety studies are a regulatory and ethical requirement prior to any testing in human subjects. Understandably, the use of in vivo models in the drug discovery process is a matter of ongoing review and debate within the scientific community and wider society [1,2]. Within this context, it is the responsibility of researchers to ensure that the in vivo models they use are justifiable ethically and designed to prosecute critical pathological mechanisms robustly while minimizing animal suffering and wastage. Moreover, particularly in an era of constrained research funding, the cost/benefit implications of each model used must be carefully considered.

Increasingly, advances in ex vivo and explant culture models using cells and tissues of human origin present credible alternatives to researchers for modelling respiratory pathology, and it is important that researcher assess whether these in vitro models better suit their research needs in comparison with established in vivo models.

It must be emphasized that none of these models is likely to be fully reflective of all aspects of human chronic respiratory diseases with their varied and often poorly understood aetiology; the best that we can hope for is that they accurately reflect pathobiological mechanisms known to be involved in the disease concerned, or sub-types thereof (Figure 1). Increasingly, these decisions will be overseen by local ethical committees.

In vivo and in vitro models of disease

Figure 1
In vivo and in vitro models of disease

The mouse is the most commonly used in vivo model of disease (A) due to the ability to selectively modulate gene expression and the need for low amounts of drugs to be used. However, other animals such as guinea pigs may be more suitable when studying cough, the function of β-agonists or eicosanoids (B). Although similar to man, the ethics and cost of non-human primates severely limits their use (C). In addition, there are a number of different in vitro approaches that are used to study respiratory diseases including precision cut lung slices (D) which are ideal for examining airway contractility, primary airway epithelial or smooth muscle cell cultures (E) grown at air–liquid interface or in submerged culture or the use of biopsies or biospheres (F) obtained from 3D culture of epithelial basal cells. The use of primary cells and tissues is important as they potentially reflect local changes occurring at the site of disease. However, they are difficult to obtain and provide limited amounts of sample.

Figure 1
In vivo and in vitro models of disease

The mouse is the most commonly used in vivo model of disease (A) due to the ability to selectively modulate gene expression and the need for low amounts of drugs to be used. However, other animals such as guinea pigs may be more suitable when studying cough, the function of β-agonists or eicosanoids (B). Although similar to man, the ethics and cost of non-human primates severely limits their use (C). In addition, there are a number of different in vitro approaches that are used to study respiratory diseases including precision cut lung slices (D) which are ideal for examining airway contractility, primary airway epithelial or smooth muscle cell cultures (E) grown at air–liquid interface or in submerged culture or the use of biopsies or biospheres (F) obtained from 3D culture of epithelial basal cells. The use of primary cells and tissues is important as they potentially reflect local changes occurring at the site of disease. However, they are difficult to obtain and provide limited amounts of sample.

Prompted by the 2nd Cross Company Respiratory Symposium (CCRS), held in Horsham in 2012, the present review aims to critique the current ‘state-of-the-art’ in the most commonly used in vivo models for respiratory research, focusing on asthma, chronic obstructive pulmonary disease (COPD) and lung fibrosis. We will highlight the strengths and weaknesses of current models, emphasizing recent improvements, and areas requiring development. In particular, we will examine the alignment of current models with evolving clinical understanding, considering aspects such as biomarker strategies, group stratification and phenotyping, modelling exacerbations and measuring quality of life impairment. Where possible, we will consider in vitro and explant models as potential replacement strategies for animal usage. The intent of the review is not to be exhaustive in referring to all the (enormous) literature, but it will refer to expert reviews and to particular papers pertinent to the discussions, especially those discussed at the CCRS and subsequently published in Clinical Science. Some of the work cited herein has previously been cited in abstract form in a published meeting report of the 2nd CCRS [3].

The ultimate aim of the present review is to promote and facilitate the discovery of new medicines for the treatment of respiratory diseases, while minimizing the time, effort and animals required to achieve this.

CHOICE OF SPECIES FOR TRANSLATIONAL MODELS OF LUNG DISEASES

There is no single ideal species for modelling any of the significant respiratory diseases; all have drawbacks and their limitations are often significant, as summarized in Table 1. Clearly randomized, blinded, controlled clinical trials in humans with the relevant disease are the gold standard, but have substantial limitations in the numbers and type of trials that can be carried out for ethical and cost reasons, as well as the nature of interventions that are feasible. We are not yet in the age of human gene knockout in vivo experiments! This can be offset to some degree by using human in vitro systems such as well-differentiated primary cell cultures with one or more lung cell types, biopsy explants and lung fragments or slices (see below) but these cannot capture the properties of the integrated system with its neuronal, endocrine and cardiovascular inputs.

Table 1
Species differences relevant to frequently used respiratory models
SpeciesFeatures and advantagesDisadvantages
Human The target species for therapy
Chronic respiratory diseases develop over years
Large body of data 
Heterogeneous, e.g. with different forms of asthma having different pathological features and responses to therapy 
  Ethical, financial, regulatory and practical constraints limit what can be done in clinical studies 
  Tissues from the affected lung not readily available 
Mouse Relatively easy and rapid breeding and husbandry
Availability of transgenics and knockouts 
Some significant differences from human lung anatomy (e.g. small size, fewer bifurcations) 
 Availability of reagents
Large body of data in mice already available 
Some significant differences from human biology/pharmacology (e.g. lack of β2-adrenergic receptors on mouse airway smooth muscle) 
  Significant sequence differences from human DNA/RNA/protein 
  Difficulty in extrapolation to human inhaled dose from mouse data 
  Model dependence on specific strains and source; do not cough 
Rat Greater availability of reagents than for non-rodent species
Better than mice for the study of inhaled agents (intratracheal dosing, dose prediction) 
Relative non-availability of transgenics and knockouts
Somewhat restricted availability of reagents (<mouse, >other species)
Airway sensitization and responses to allergen (especially LAR) can vary over time and between laboratories and animal source 
Guinea pig Lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes
Greater similarities with human lung anatomy and pharmacology than mouse
Do cough and sneeze 
Non-availability of transgenics and knockouts
Restricted availability of reagents
Airway sensitization and responses to allergen (especially LAR) can vary over time and between laboratories and animal source 
Sheep, dog Amenable to detailed study of respiratory and cardiovascular systems
Outbred 
Restricted availability of reagents; amounts of reagent needed
Data from few laboratories only, reputation that ‘everything works’? 
Non-human primate Marmoset, Cynomolgous monkey, Rhesus monkey Expensive and relatively slow, similar to clinical trials 
 In principle closer to human: fewer sequence differences from human DNA/RNA/protein
Higher degree of cross-reactivity with human receptors/antibodies: appropriate for the study of many biologicals
Husbandry easier with marmosets than large primates 
Ethical considerations and constraints
Less validation data available than with some other species 
SpeciesFeatures and advantagesDisadvantages
Human The target species for therapy
Chronic respiratory diseases develop over years
Large body of data 
Heterogeneous, e.g. with different forms of asthma having different pathological features and responses to therapy 
  Ethical, financial, regulatory and practical constraints limit what can be done in clinical studies 
  Tissues from the affected lung not readily available 
Mouse Relatively easy and rapid breeding and husbandry
Availability of transgenics and knockouts 
Some significant differences from human lung anatomy (e.g. small size, fewer bifurcations) 
 Availability of reagents
Large body of data in mice already available 
Some significant differences from human biology/pharmacology (e.g. lack of β2-adrenergic receptors on mouse airway smooth muscle) 
  Significant sequence differences from human DNA/RNA/protein 
  Difficulty in extrapolation to human inhaled dose from mouse data 
  Model dependence on specific strains and source; do not cough 
Rat Greater availability of reagents than for non-rodent species
Better than mice for the study of inhaled agents (intratracheal dosing, dose prediction) 
Relative non-availability of transgenics and knockouts
Somewhat restricted availability of reagents (<mouse, >other species)
Airway sensitization and responses to allergen (especially LAR) can vary over time and between laboratories and animal source 
Guinea pig Lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes
Greater similarities with human lung anatomy and pharmacology than mouse
Do cough and sneeze 
Non-availability of transgenics and knockouts
Restricted availability of reagents
Airway sensitization and responses to allergen (especially LAR) can vary over time and between laboratories and animal source 
Sheep, dog Amenable to detailed study of respiratory and cardiovascular systems
Outbred 
Restricted availability of reagents; amounts of reagent needed
Data from few laboratories only, reputation that ‘everything works’? 
Non-human primate Marmoset, Cynomolgous monkey, Rhesus monkey Expensive and relatively slow, similar to clinical trials 
 In principle closer to human: fewer sequence differences from human DNA/RNA/protein
Higher degree of cross-reactivity with human receptors/antibodies: appropriate for the study of many biologicals
Husbandry easier with marmosets than large primates 
Ethical considerations and constraints
Less validation data available than with some other species 

The mouse has been a highly favoured species for modelling lung disease, as for other diseases, for a range of good reasons (Table 1) which particularly favour their use in academic research; the use of mice does, however, have significant drawbacks. There is more discussion of mouse models and their clinical predictability or lack thereof in the specific lung diseases below.

Other species have been used for specific purposes and reasons, as outlined in Table 1. For example, primates have been used for the study of biological therapies which do not cross-react with the target proteins in other species.

TRANSLATIONAL MODELS OF ASTHMA

There have been some substantial advances in the understanding of the pathophysiology and nature of asthma over the last decade, which impact significantly on our understanding in terms of what are appropriate translational models, in vitro and in vivo. Moreover, some of the models that have been in use for many years have shown substantial limitations and failures in predicting clinical trial outcomes in asthma (Table 2, [3]).

Table 2
Asthma models ([3,5] and references therein)
ModelSpecies (strain)Trying to modelStrengths/featuresWeaknesses
Acute ovalbumin Mouse Allergic asthma Short, easy to carry out
Reproduce some features of human asthma (especially increased lung eosinophils and AHR, Th2 phenotype)
Availability of transgenics, knockouts, reagents (cytokines, chemokines, antibodies) 
Not predictive of efficacy in human asthma
Do not robustly give lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes
Some significant differences from human lung anatomy and pharmacology 
Chronic house dust mite Mouse Allergic asthma More predictive of efficacy in human asthma (so far)
Reproduce more features of human asthma (especially increased lung eosinophils and neutrophils, AHR, somewhat more mixed T-cell phenotype) 
Take 5–7 weeks to carry out
Do not robustly give lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
Chronic ovalbumin Mouse Allergic asthma More predictive of efficacy in human asthma (so far) Less widely used and characterized as yet 
   Reproduce more features of human asthma (especially increased lung eosinophils and neutrophils, AHR, somewhat more mixed T-cell phenotype) Take 5–7 weeks to carry out
Do not robustly give lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
House dust mite/
FCA sensitized 
Mouse Severe/mixed phenotype asthma Reproduce some features of human asthma (especially increased lung eosinophils and neutrophils, AHR, more mixed Th2, Th1, Th17 cell phenotype)
At least partially resistant to glucocorticoids 
Less widely used and characterized as yet
Take 3–7 weeks to carry out
Do not robustly give lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
Acute ovalbumin Guinea pig Allergic asthma Short, easy to carry out
Reproduce some features of human asthma (especially increased lung eosinophils and AHR, Th2 phenotype)
Lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
Not predictive of efficacy in human asthma
Airway sensitization and responses to allergen (especially LAR) can vary over time and between laboratories 
Chronic ovalbumin Guinea pig Allergic asthma More predictive of efficacy in human asthma (so far) Less widely used and characterized as yet 
   Reproduce some features of human asthma (especially increased lung eosinophils and AHR, Th2 phenotype)
Lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
Take weeks to carry out
Airway sensitization and responses to allergen (especially LAR) can vary over time and between laboratories 
Acute ovalbumin Rat (Brown Norway) Allergic asthma Short, easy to carry out
Reproduce some features of human asthma especially increased lung eosinophils and AHR, Th2 phenotype, lung function responses (e.g. early and late phase) to allergen 
Do not robustly give distinct early and late phase lung function responses as seen in humans
Airway sensitization and responses to allergen can vary over time and between laboratories
Non-availability of transgenics and knockouts 
Various Non-human primate Allergic asthma In principle closer to human Expensive and relatively slow, ethical considerations and constraints, less validation data available 
ModelSpecies (strain)Trying to modelStrengths/featuresWeaknesses
Acute ovalbumin Mouse Allergic asthma Short, easy to carry out
Reproduce some features of human asthma (especially increased lung eosinophils and AHR, Th2 phenotype)
Availability of transgenics, knockouts, reagents (cytokines, chemokines, antibodies) 
Not predictive of efficacy in human asthma
Do not robustly give lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes
Some significant differences from human lung anatomy and pharmacology 
Chronic house dust mite Mouse Allergic asthma More predictive of efficacy in human asthma (so far)
Reproduce more features of human asthma (especially increased lung eosinophils and neutrophils, AHR, somewhat more mixed T-cell phenotype) 
Take 5–7 weeks to carry out
Do not robustly give lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
Chronic ovalbumin Mouse Allergic asthma More predictive of efficacy in human asthma (so far) Less widely used and characterized as yet 
   Reproduce more features of human asthma (especially increased lung eosinophils and neutrophils, AHR, somewhat more mixed T-cell phenotype) Take 5–7 weeks to carry out
Do not robustly give lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
House dust mite/
FCA sensitized 
Mouse Severe/mixed phenotype asthma Reproduce some features of human asthma (especially increased lung eosinophils and neutrophils, AHR, more mixed Th2, Th1, Th17 cell phenotype)
At least partially resistant to glucocorticoids 
Less widely used and characterized as yet
Take 3–7 weeks to carry out
Do not robustly give lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
Acute ovalbumin Guinea pig Allergic asthma Short, easy to carry out
Reproduce some features of human asthma (especially increased lung eosinophils and AHR, Th2 phenotype)
Lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
Not predictive of efficacy in human asthma
Airway sensitization and responses to allergen (especially LAR) can vary over time and between laboratories 
Chronic ovalbumin Guinea pig Allergic asthma More predictive of efficacy in human asthma (so far) Less widely used and characterized as yet 
   Reproduce some features of human asthma (especially increased lung eosinophils and AHR, Th2 phenotype)
Lung function responses (e.g. early and late phase) to allergen equivalent to human FEV1 changes 
Take weeks to carry out
Airway sensitization and responses to allergen (especially LAR) can vary over time and between laboratories 
Acute ovalbumin Rat (Brown Norway) Allergic asthma Short, easy to carry out
Reproduce some features of human asthma especially increased lung eosinophils and AHR, Th2 phenotype, lung function responses (e.g. early and late phase) to allergen 
Do not robustly give distinct early and late phase lung function responses as seen in humans
Airway sensitization and responses to allergen can vary over time and between laboratories
Non-availability of transgenics and knockouts 
Various Non-human primate Allergic asthma In principle closer to human Expensive and relatively slow, ethical considerations and constraints, less validation data available 

The animal models of asthma that have for some time been developed and utilized to model asthma have relied on sensitization to and subsequent airway challenge with allergens, often the model antigen ovalbumin, resulting in pathobiological changes which appeared to strongly resemble and recapitulate allergic asthma in humans. These changes include elevated levels of IgE, release of mediators, influx of inflammatory cells especially eosinophils into the airways, goblet cell hyperplasia, epithelial hypertrophy, bronchoconstriction, airway hyper-responsiveness (AHR) to bronchoconstrictors (Table 2, reviewed in [1,46]. This pattern of human-like responses, together with appropriate beneficial effects of glucocorticoids and beta2 receptor agonists in these models, led to a presumption that beneficial effects of novel medicinal mechanisms in such models would predict efficacy in human asthma. However, this hope has been dashed by a series of failures to translate.

One specific well-studied example is that of the development and testing of a selective inhibitor of the inducible form of nitric oxide synthase (iNOS, NOS-2), GW274150. This potent and highly selective inhibitor was identified and developed and shown to have high oral bioavailability and an appropriate half-life for oral, once-daily dosing [7]. GW274150 was shown to effectively inhibit iNOS in vivo, and was effective in the acute ovalbumin sensitization and challenge models in wide use at the time, in both mice and guinea pigs, decreasing the eosinophil influx into the lung and the AHR in mice, as well as substantially suppressing the late phase bronchoconstriction response (LAR) resulting from allergen challenge in guinea pigs. Phase 1 and 2 studies established that GW274150 substantially inhibited iNOS in humans; it was very disappointing therefore when a clinical allergen challenge study showed no benefits of inhibition of iNOS with GW274150 despite a good matching of end points: no effects on sputum eosinophils, early or late phase bronchoconstriction responses or AHR [8]. Subsequent studies have shown that chronic animal models appear to provide a better correlation with the clinical findings: in both mice [5] and guinea pigs [9], GW274150 loses its beneficial effects in chronic models in contrast with the benefits seen in acute models. In this case, the failure to predict clinical outcomes from pre-clinical studies appears to be related to the chronicity of the human condition not being well represented by acute animal models. There is substantial work published and ongoing to develop, refine and further characterize such models, such as the 5–7 week house dust mite (HDM) allergen model (e.g. [1,5,10], including work using non-invasive methods to monitor changes in the lung [11].

In other models, there may be additional factors in play, e.g. a simple allergen such as ovalbumin may not adequately model the effects of complex allergens associated with allergic asthma, such as house-dust mite, pollens and animal dander. These latter models all have components other than the sensitized allergen, including proteases, chitins and lipopolysaccharide (LPS) which are likely to change the pathophysiological responses and the effects on potential therapies.

Another factor which may play a role in poor pre-clinical to clinical prediction is where the pre-clinical end points do not match well what is to be measured clinically. For example, some eosinophil-directed therapies [such as the anti-interleukin (IL)-5 antibody mepolizumab] succeeded in decreasing lung eosinophil levels but failed to provide lung function benefits in clinical studies [12]. Subsequent work has shown that there is a clinical benefit of mepolizumab, but it is in decreasing the number of exacerbations suffered, an end point that was not studied in the original pre-clinical work. Models of exacerbation of asthma are under development but are by no means established even now (see discussion below and Table 4).

Furthermore, it can be particularly crucial to select the most appropriate species (Table 1) for studies of particular mechanisms in asthma. For example, mice (and the pig, cat and rabbit) have differences in the function and relative importance of β1- vs β2-adrenergic receptors in the airways (β1 predominant) as compared with human and guinea pig airways (β2 predominant, [13]; in addition, the repertoire of mediators released by mouse and rat mast cells is also different from guinea pigs and humans [14], and this may explain at least in part why distinct early and late phase allergic airway responses are not easily modelled in mice and rats). For this reason, it seems important that other species are considered where there may be such a discrepancy between, in particular, mouse and human biology. This could be guinea pigs, or non-rodents, including non-human primates (NHPs) such as marmosets (Table 1, [15]), cynomolgous or rhesus monkeys. These NHPs may be the only option for modelling the effects of biologicals or other agents which do not cross-react with non-primate species, but there are clearly ethical, cost and practical constraints in their use, and there is much less data characterizing them (e.g. chronic allergen exposure models such as those increasingly used in rodents) or validating their clinical predictability with agents known to be effective or ineffective in human asthma.

Matching of pre-clinical models to sub-populations of human asthma seems likely to be important (Figure 2) as again reflected in the experience with mepolizumab: IL-5 and eosinophils are unlikely to play a significant role in non-eosinophilic asthma, such that outcomes in eosinophilic animal models are only likely to translate to eosinophilic human asthma, and not to non-eosinophilic asthma.

Matching asthma phenotypes to models of disease

Figure 2
Matching asthma phenotypes to models of disease

Numerous animal models of asthma exist each with their own specific characteristics. The ability to predict drug efficacy will improve when we map distinct clinical or molecular phenotypes to the correct animal model of disease. For example, studies investigating drugs that will be used to treat a patient with neutrophilic Th17 asthma should be used in the appropriate model and not a model that is Th2-driven for example.

Figure 2
Matching asthma phenotypes to models of disease

Numerous animal models of asthma exist each with their own specific characteristics. The ability to predict drug efficacy will improve when we map distinct clinical or molecular phenotypes to the correct animal model of disease. For example, studies investigating drugs that will be used to treat a patient with neutrophilic Th17 asthma should be used in the appropriate model and not a model that is Th2-driven for example.

Translational models of severe asthma are a challenge, not only because of this heterogeneity but also because they are (by definition) uncontrolled despite substantial therapy with high dose inhaled corticosteroids (ICS), and in many cases also oral corticosteroids, implying a failure of these powerful anti-inflammatories to fully control the disease. This has led to a search for models of asthma that are at least in part resistant to corticosteroids, both in vivo (e.g. [16]) and in vitro.

There is significant interest in trying to address the unmet clinical needs of children with asthma, but we are however, even further behind in identifying predictive models of paediatric asthma.

Finally, there have been demonstrated to be substantial differences between strains of rodents used for models of allergic asthma, between genders and even between laboratories apparently using the same protocols. Even within laboratories, these models can be inconsistent over time or depending on individual batches of reagents. However, the chronic HDM and the Freund's complete adjuvant (FCA)-HDM models have proved to be consistent between laboratories and over time so far, as reported at the cross-company symposium, at least when using standardized, shared protocols within the Unbiased BIOmarkers in PREDiction of respiratory disease outcomes project (U-BIOPRED) consortium.

Thus, there appear to be several important factors in these failures to translate from animal models to human asthma: (i) acute (typically <3 weeks) versus chronic (≥5 weeks) models; (ii) simple versus complex allergens/other challenges used; (iii) prophylactic versus therapeutic dosing; (iv) end points not matching clinical ones; (v) heterogeneity within the human asthmatic population; (vi) strain/gender/laboratory differences; and (vii) inappropriate choice of species.

There is of course a great desire to use in vitro models instead of in vivo animal models where possible. This has led to significant work using cells, tissues such as lung slices and explants, from human subjects wherever possible; some of these systems are summarized and discussed below.

ANIMAL MODELS OF COPD

The range of animal models of COPD is indicated in Figure 3 and Table 3. COPD models do not usually recapitulate all the major features of human COPD and are commonly based on the induction of selected COPD-like lesions in the lungs and airways using noxious inhalants such as tobacco smoke (TS), ozone or sulfur dioxide or exposure to proteases [6,1725]. Depending on the duration and intensity of exposure, these noxious stimuli induce signs of chronic inflammation, small airways disease, bronchitis, emphysema and frequent exacerbations in a highly species- and strain-dependent manner (Figure 3, [17]). In addition, transgenic and knockout mice often spontaneously develop COPD-like lesions with emphysema providing molecular insights into pathophysiological mechanisms. It is imperative, therefore, to design and use models that will answer the precise question being asked whether it relate to pathophysiology or drug response [26]. However, current models often fail to mimic irreversible airflow obstruction with associated with cough and sputum production and future developments should aim to address this. Increased use of prophylactic interventions would also help to increase the predictive nature of these models [6,1820,2225,2729]. The ability to obtain clear exacerbations of inflammation and airway function will also need to be addressed.

Table 3
COPD models
ModelSpecies (strain)Trying to modelStrengths/featuresWeaknesses
All   Most current models designed to induce emphysema Rarely mimic small airways disease, bronchitis and exacerbations 
   Mild form of centrilobular emphysema/small airway remodelling and changes in airway function akin to GOLD stage I/II Many studies report prophylactic rather than therapeutic interventions
Patients are treated as they progress through GOLD II/III/IV 
Tobacco smoke: nose-only Mouse COPD
Inflammation 
Better control of dosage and exposure Whole body restraint is a known stressor for mice 
  Emphysema Targeted delivery to the respiratory system  
   Combination of mainstream and side stream smoke  
Tobacco smoke:
whole body 
Mouse COPD
Inflammation
Emphysema 
Does not require restraint
Less labour intensive
Combination of mainstream and side stream smoke 
TS exposure is not as well regulated as nose-only TS exposure
animals can also ingest deposits of nicotine and tar from the fur of exposed animals 
Tobacco smoke (acute) Mouse COPD
Inflammation 
Uses major causative factor in human COPD
Inflammatory changes 
No airway remodelling or emphysema until at least 6 months exposure
Differ considerably in respiratory tract function and anatomy from man 
Tobacco smoke (chronic) Mouse COPD
Emphysema 
Uses major causative factor in human COPD
Results in emphysema-like changes in lung
BAL inflammation and lung function are steroid insensitive
Inflammation, emphysema and mucus production are self-sustaining
Links to chronicity and age-associated aspects of COPD 
Takes months of TS exposure and no standard protocol
Long-term exposure models do not show large changes in lung function making observations of drug effects difficult
Differ considerably in respiratory tract function and anatomy from man
Difficult to assess lung function in mice 
Tobacco smoke (8 week intense) Mouse COPD
Inflammation
Emphysema 
Chronic airway/lung inflammation, mucus hypersecretion, emphysema-like changes and reduced lung function
Similar effect in C57BL/6 and BALB/c strains
Mice more susceptible to infection 
Labour intensive: twelve 3R4F reference cigarettes twice per day for 75 min and 5 times per week for 8 weeks using nose-only TS exposure 
Tobacco smoke (acute and chronic) Rat COPD
Inflammation
Emphysema 
Uses major causative factor in human COPD Relatively resistant to the induction of emphysema-like lesions in a strain-dependent manner 
Ozone (6 weeks) Mouse and rat COPD
Inflammation
Emphysema 
Lung function changes similar to that seen in human COPD
Increases mean linear intercept (Lm) and AHR to acetylcholine
Changes persist after cessation of exposure 
Comparison with smoke-induced changes unclear 
Ultrafine particles Mouse and rat COPD
Emphysema 
Include silica, coal dust and diesel exhaust particles (DEP) Predominantly characterized by focal emphysema 
   Increase oxidative stress  
   Similar mechanisms to smoke-induced emphysema  
   Provide link with tumorigenesis  
Tissue-degrading enzymes Mouse and rat COPD
Emphysema 
For example, neutrophil elastase, porcine pancreatic elastase and papain
Panacinar emphysema, inflammatory cell influx into the lungs; and systemic inflammation–akin to α1 anti-trypsin disease 
Lack of knowledge regarding the differences in injury pathways compared with smoking models 
   Used to study mechanisms related to emphysema and lung repair  
LPS Rat or mouse COPD
Inflammation 
Neutrophilic and macrophage inflammation with concomitant airspace enlargement Endotoxin-induced neutrophilic inflammation is corticosteroid sensitive 
Transgenic animals Mouse COPD
Inflammation
Emphysema 
Conditional knockouts and knockin models enable the distinction between developmental airspace enlargement and adult emphysema Data confusing due to the use of unconditional knockouts and knockins 
Tobacco smoke (acute/chronic) Rat COPD
Inflammation
Emphysema 
 Relatively resistant to the induction of emphysema-like lesions in a strain-dependent manner 
Tobacco smoke (chronic) Guinea pigs COPD
Inflammation
Emphysema 
Develop COPD-like features earlier than mice
Pathology of airway closer to that of man
Emphysema stable after smoking cessation 
Few tools available for molecular analysis
Lack of transgenic animals and the limited numbers of species strains
A prominent axon reflex 
SO2 (daily exposure for 4–8 weeks) Rat COPD
Inflammation
Emphysema 
Tissue destruction similar to that seen in end-stage emphysema
Neutrophilic inflammation, increased mucus production and mucus cell metaplasia and damage of ciliated epithelial cells 
Neutrophilic inflammation is steroid sensitive 
Large animal models All species COPD
Inflammation
Emphysema 
Good match in gross anatomy, morphometry and physiology of their respiratory systems Costs, lack of reagents
Reduced understanding of disease mechanisms in relation to that in rodents 
Inhaled neutrophil elastase Sheep Mucociliary function  No TS exposure model 
Tobacco smoke, SO2 or proteolytic enzymes Dogs Chronic bronchitis
Emphysema 
Similar pathology and pathophysiology to man
Cough reflex, mucus production, collagen deposition and bronchoconstriction
Repeat bronchoscopies possible 
This model is difficult to work with in relation to regularity issues and to the high costs involved
Lack of molecular tools Heterogeneity of animals 
Tobacco smoke Non-human primates COPD
Inflammation
Emphysema 
In principle closer to human
Effects of maternal smoke exposure during pregnancy
Carcinogenic effects of TS 
Regularity/ethical issues
High costs 
ModelSpecies (strain)Trying to modelStrengths/featuresWeaknesses
All   Most current models designed to induce emphysema Rarely mimic small airways disease, bronchitis and exacerbations 
   Mild form of centrilobular emphysema/small airway remodelling and changes in airway function akin to GOLD stage I/II Many studies report prophylactic rather than therapeutic interventions
Patients are treated as they progress through GOLD II/III/IV 
Tobacco smoke: nose-only Mouse COPD
Inflammation 
Better control of dosage and exposure Whole body restraint is a known stressor for mice 
  Emphysema Targeted delivery to the respiratory system  
   Combination of mainstream and side stream smoke  
Tobacco smoke:
whole body 
Mouse COPD
Inflammation
Emphysema 
Does not require restraint
Less labour intensive
Combination of mainstream and side stream smoke 
TS exposure is not as well regulated as nose-only TS exposure
animals can also ingest deposits of nicotine and tar from the fur of exposed animals 
Tobacco smoke (acute) Mouse COPD
Inflammation 
Uses major causative factor in human COPD
Inflammatory changes 
No airway remodelling or emphysema until at least 6 months exposure
Differ considerably in respiratory tract function and anatomy from man 
Tobacco smoke (chronic) Mouse COPD
Emphysema 
Uses major causative factor in human COPD
Results in emphysema-like changes in lung
BAL inflammation and lung function are steroid insensitive
Inflammation, emphysema and mucus production are self-sustaining
Links to chronicity and age-associated aspects of COPD 
Takes months of TS exposure and no standard protocol
Long-term exposure models do not show large changes in lung function making observations of drug effects difficult
Differ considerably in respiratory tract function and anatomy from man
Difficult to assess lung function in mice 
Tobacco smoke (8 week intense) Mouse COPD
Inflammation
Emphysema 
Chronic airway/lung inflammation, mucus hypersecretion, emphysema-like changes and reduced lung function
Similar effect in C57BL/6 and BALB/c strains
Mice more susceptible to infection 
Labour intensive: twelve 3R4F reference cigarettes twice per day for 75 min and 5 times per week for 8 weeks using nose-only TS exposure 
Tobacco smoke (acute and chronic) Rat COPD
Inflammation
Emphysema 
Uses major causative factor in human COPD Relatively resistant to the induction of emphysema-like lesions in a strain-dependent manner 
Ozone (6 weeks) Mouse and rat COPD
Inflammation
Emphysema 
Lung function changes similar to that seen in human COPD
Increases mean linear intercept (Lm) and AHR to acetylcholine
Changes persist after cessation of exposure 
Comparison with smoke-induced changes unclear 
Ultrafine particles Mouse and rat COPD
Emphysema 
Include silica, coal dust and diesel exhaust particles (DEP) Predominantly characterized by focal emphysema 
   Increase oxidative stress  
   Similar mechanisms to smoke-induced emphysema  
   Provide link with tumorigenesis  
Tissue-degrading enzymes Mouse and rat COPD
Emphysema 
For example, neutrophil elastase, porcine pancreatic elastase and papain
Panacinar emphysema, inflammatory cell influx into the lungs; and systemic inflammation–akin to α1 anti-trypsin disease 
Lack of knowledge regarding the differences in injury pathways compared with smoking models 
   Used to study mechanisms related to emphysema and lung repair  
LPS Rat or mouse COPD
Inflammation 
Neutrophilic and macrophage inflammation with concomitant airspace enlargement Endotoxin-induced neutrophilic inflammation is corticosteroid sensitive 
Transgenic animals Mouse COPD
Inflammation
Emphysema 
Conditional knockouts and knockin models enable the distinction between developmental airspace enlargement and adult emphysema Data confusing due to the use of unconditional knockouts and knockins 
Tobacco smoke (acute/chronic) Rat COPD
Inflammation
Emphysema 
 Relatively resistant to the induction of emphysema-like lesions in a strain-dependent manner 
Tobacco smoke (chronic) Guinea pigs COPD
Inflammation
Emphysema 
Develop COPD-like features earlier than mice
Pathology of airway closer to that of man
Emphysema stable after smoking cessation 
Few tools available for molecular analysis
Lack of transgenic animals and the limited numbers of species strains
A prominent axon reflex 
SO2 (daily exposure for 4–8 weeks) Rat COPD
Inflammation
Emphysema 
Tissue destruction similar to that seen in end-stage emphysema
Neutrophilic inflammation, increased mucus production and mucus cell metaplasia and damage of ciliated epithelial cells 
Neutrophilic inflammation is steroid sensitive 
Large animal models All species COPD
Inflammation
Emphysema 
Good match in gross anatomy, morphometry and physiology of their respiratory systems Costs, lack of reagents
Reduced understanding of disease mechanisms in relation to that in rodents 
Inhaled neutrophil elastase Sheep Mucociliary function  No TS exposure model 
Tobacco smoke, SO2 or proteolytic enzymes Dogs Chronic bronchitis
Emphysema 
Similar pathology and pathophysiology to man
Cough reflex, mucus production, collagen deposition and bronchoconstriction
Repeat bronchoscopies possible 
This model is difficult to work with in relation to regularity issues and to the high costs involved
Lack of molecular tools Heterogeneity of animals 
Tobacco smoke Non-human primates COPD
Inflammation
Emphysema 
In principle closer to human
Effects of maternal smoke exposure during pregnancy
Carcinogenic effects of TS 
Regularity/ethical issues
High costs 

Matching COPD phenotypes to models of disease

Figure 3
Matching COPD phenotypes to models of disease

Numerous animal models of COPD exist each with their own specific characteristics. The ability to predict drug efficacy will improve when we are able to map distinct clinical or molecular phenotypes to the correct disease model. For example, studies investigating drugs that will be used to treat a patient with small airways disease should be used in the appropriate model as the results may not be translated from a model where emphysema is the major end point.

Figure 3
Matching COPD phenotypes to models of disease

Numerous animal models of COPD exist each with their own specific characteristics. The ability to predict drug efficacy will improve when we are able to map distinct clinical or molecular phenotypes to the correct disease model. For example, studies investigating drugs that will be used to treat a patient with small airways disease should be used in the appropriate model as the results may not be translated from a model where emphysema is the major end point.

Combining models of inhaled exposure, proteinase-based tissue degradation and gene-targeting approaches may encompass more features of the disease, it may be that these models will be less effective at modelling human disease due to patient heterogeneity. As such, a model that only produces a single pathologic or phenotypic COPD feature may be more useful for the testing of new chemical entities (NCEs) or biologics designed to target that disease feature [6,1720,2225,27].

TS exposure can occur via nose-only or whole body systems [17,19,21,25]. An advantage of nose-only TS exposure is the better control of dosage and exposure as well as targeted delivery to the respiratory system [17,19,24]. In contrast, whole body TS exposure does not require restraint, animals have free access to food and water and the system is less labour intensive. However, whole body TS exposure is not as well regulated as nose-only TS exposure and animals can also ingest deposits of nicotine and tar from the fur of exposed animals [17,19,24]. Both systems deliver a combination of mainstream and side stream smoke [19,20]. Currently there is no standard protocol for these TS models and they vary in delivery systems, species or strains, cigarette makes, main-stream versus side-stream smoke and varied smoke dose/exposure times [6,21]. This can make extrapolating findings from one group to the next difficult [6]. Importantly, bronchoalveolar lavage (BAL) inflammation and lung function in both acute and chronic TS exposure models are insensitive to both oral and ICS at doses which induce weight loss and suppress the inflammatory response in LPS-induced neutrophilic models of COPD [18].

A variety of animal species has been exposed to TS including guinea pigs, rabbits, dogs, rats and mice. Guinea pigs have been reported to be a very susceptible species as they develop COPD-like lesions and emphysema-like airspace enlargement within a few months of active TS exposure [20]. In addition, the β2- and muscarinic receptor pharmacology is almost identical with that seen in man [30]. There are clear mouse-strain-dependent effects of smoking on the susceptibility to emphysema and inflammatory responses and this may also vary according to the smoke delivery system [19]. The use of advanced imaging techniques may provide better approaches for determining the efficacy of some drugs targeting small airways and emphysema [23].

Mice offer the opportunity to manipulate gene expression, however, it is more difficult to assess lung function in mice and they differ considerably in respiratory tract function and anatomy compared with humans: they are obligate nose breathers, they have lower numbers of cilia, fewer Clara cells and a restriction of submucosal glands to the trachea [20]. Importantly, mice do not have a cough reflex and many mediators such as histamine or tachykinins have different pharmacological effects in mice compared with man. Therefore, the selection of a strain needs to be done with great caution.

A recent 8-week TS exposure model has been developed which reproduces many of the major pathological features of disease; chronic airway/lung inflammation, mucus hypersecretion, emphysema-like changes and reduced lung function which is short enough in duration to be amenable for drug intervention [31]. Mice undergoing this regime also suffer more severe Streptococcus pneumoniae, Haemophilus influenzae and influenza infections than air exposed animals [31].

Exposure to ozone and particulates, as well as the use of lung proteases have also been used as models of COPD. Chronic (6 week) ozone exposure in mice causes significant lung injury with some features related to inflammatory and emphysematous changes occurring in human COPD as well as changes in lung function and AHR [20,32,33]. Exposure to ultrafine particles, silica, coal dust and diesel exhaust particles (DEP) has been used which all cause significant levels of oxidative stress, inflammation and emphysema [20]. Emphysema-like lesions can also be achieved by intrapulmonary challenge with tissue-degrading enzymes such as human neutrophil elastase, porcine pancreatic elastase, or papain [20]. These latter models are best used to study mechanisms specifically related to emphysema and to the repair of damaged lung [18,34].

Although the use of invasive and non-invasive lung function techniques, as well as new imaging techniques, can provide useful data on mouse pulmonary function [11], the small size of the mouse and its respiratory system presents some problematic issues when measuring the pathophysiologic parameters of airway disease [23]. As a result, there has been some effort to use larger animals such as dogs, sheep and monkeys as models of a number of respiratory diseases including COPD [35]. Large animals have a better match in gross anatomy including presence of bronchial glands, goblet cells and ciliated epithelial cells; morphometry; structure including the ability to remodel the bronchial circulation; and physiology of their respiratory systems compared with man [35]. Sheep have been used to study mucociliary function in COPD after inhaled human neutrophil elastase [36] but there are no published large non-primate TS models except for early data in dogs where repeated bronchoscopies have been performed to examine immune and inflammatory readouts [35,37]. NHP smoking models have also been used primarily for studying the effects of maternal smoke exposure during pregnancy as well as the carcinogenic effects of cigarette smoke [19,37,38]. This model is difficult to work with in relation to regularity issues and to the high costs involved.

In future, bioinformatic comparisons of signature genes or pathways from clinical specimens from patients with different sub-phenotypes of COPD need to be integrated with clinical observations and similar phenotypic end points as clinical studies and related back to in vivo models in order to further understand, define and treat COPD (Figure 3, [19,39,40]). Only then can animal models be tailored precisely to the specific physiological and pharmacological features of human COPD which each model possesses [6,1720,2225,32,33].

TRANSLATIONAL MODELS OF LUNG FIBROSIS

Idiopathic pulmonary fibrosis (IPF) is an insidious progressive condition of appalling prognosis, with a median survival of 3 years from diagnosis. Currently there is a paucity of therapies, in part due to our incomplete understanding of key pathological mechanisms (reviewed in [41]).

The classical histopathological pattern of IPF–usual interstitial pneumonia (UIP)–is characterized by evidence of patchy parenchymal epithelial damage, type II hyperplasia, together with abnormal proliferation and activation of fibroblasts into highly synthetic α-smooth muscle actin (αSMA)–positive myofibroblasts and extensive and disorganized deposition of collagen and other extracellular matrix (ECM) proteins [41]. The gradual obliteration of functional alveolar units eventually culminates in respiratory failure and death of affected individuals. The underlying mechanisms driving these processes remain unclear however current theories propose that IPF develops as a result of an aberrant wound healing response to repeated micro-injury. Tissue repair and developmental signatures dominate microarray studies of human disease [42,43], and a variety of dysregulated wound healing responses including defective epithelial reconstitution and crosstalk with mesenchymal cells, uncontrolled coagulation, abnormal neovascularization, and inadequate antioxidant control have been postulated as contributing to pathology [44]. These dysfunctional processes appear to be maintained by a complex network of mitogenic factors and morphogens, most prominently transforming growth factor β (TGFβ) as well as aberrant cellular interactions with degenerate ECM [45].

While initiating factors in IPF remain enigmatic, the fibro-proliferative wound healing responses to a variety of injurious lung insults which elicit an inflammatory response have been exploited to study fibrotic disease mechanism. The bleomycin model is the most commonly utilized of these and will form the bulk of the present review, however other models including, radiation induced damage, and instillation of FITC, silica or asbestos injury have been established. These have been thoroughly reviewed previously [4648] with salient features listed in Table 4. Common to these models is their reliance on the fibro-proliferative response to inflammation, with the implicit risk for drug discovery that anti-inflammatory interventions may show false anti-fibrotic efficacy unless adequately controlled for. Distinct from inflammatory based insults, models based around specific pathological mechanisms implicated in pulmonary fibrosis include over-expression pro-fibrotic cytokines {active-TGFβ, IL-1β and tumour necrosis factor (TNF) α [4951]}, transgenic over expression of pro-fibrotic cytokines (TGFα, TGF-β and IL-13 [5255]), intravenous instillation of human fibroblasts into immunodeficient mice non-obese diabetic (NOD)–SCID mice [56,57], and targeted injury of type II alveolar epithelial cells [58] (also see Table 4). The major criticisms of all these models are that none completely recapitulate all of the histological features of IPF, lacking progressive disease pathology, and often showing spontaneous reversibility of lesions. Although reconstituting all pathological features within a single model could be considered the Holy Grail, it must be noted that this is difficult to achieve in an economically viable format which maintains experimental reproducibility. It is noteworthy that spontaneously occurring fibrotic disorders in larger animals bear striking similarities to human disease, most notably West Highland terriers [59] and donkeys [60]. Although these are of interest, it remains unclear whether their reproducibility and cost would be amenable to the drug discovery process.

Table 4
Fibrosis models ([4648] and references therein)

Abbreviations: d=day.

ModelSpeciesFeatures/practicalitiesStrengthsWeaknesses
Bleomycin Mouse Direct cellular injury due to DNA scission
Inflammatory infiltration (d0–7 in single dose model, more persistent in multiple dose model) and vascular leak followed by fibrotic phase
Multiple delivery routes intratracheal, oropharyngeal: (Single dose model: d14–28
Multiple dose: 16 weeks
Other routes: i.v., i.t., i.p. 
Single dose model very well characterized model
Clinically relevant (bleomycin toxicity)
Rapid time frame for development and resolution (reports of resolution are conflicting)
Multiple dose model (intratracheal)
Mimics clinically relevant repetitive lung injury aetiology of PF
Prominent AEC hyperplasia, robust fibrosis similar to human pathology 
Single dose model (i.t./o.p.): Significant inflammatory involvement in early phases (neutrophils, d0–7)
Self-limiting without chronic progression (d28+, reports of resolution are conflicting)
Lack of alveolar epithelial cell hyperplasia seen in human disease
Strain specific (C57Bl6> SV> Balbc) Lot variability of bleomycin activity
Multiple dose model (i.t.): long duration of model (16 weeks), late attenuation of neutrophilia 
FITC Mouse As above Ability to visualize areas of lung injury by characteristic green fluorescence
Time frame for development of fibrosis is d14–28
Fibrotic response persists for >6 months
Not strain dependent 
Not clinically relevant
Lot variability of FITC
FITC must be made fresh and is subject to user variability 
TGFβ overexpression Rat Adenoviral delivery or doxycycline regulated transgene expression in epithelium
TGFβ expression (d1–7)
Early mononuclear infiltration (d3–7)
Alveolar consolidation, epithelial apoptosis and collagen accumulation 
Mimics human disease
Enables single growth
Factor dissection of pathology
Persistent fibrotic injury 
Early inflammatory phase due to adenoviral vector
Strain dependence C57Bl6> Balb/c 
Irradiation Mouse Free radical mediated DNA damage
Induction of inflammation followed by fibrosis
Vascular remodelling reminiscent of PAH 
Clinically relevant
Low mortality 
Long duration of model
Need to protect other organs from irradiation
Expensive
Strain dependence: C57Bl6> C3H/HeJ and CBA/J (resistant) 
Silica Mouse Silica retained in lung inducing inflammatory response and development of fibrotic nodules around deposits
Characterized by activation of the inflammasome activation
Multiple routes of administration: aerosol (d40–120), i.t./o.p. (d14–28)
Baking of silica required to destroy endotoxins 
Clinically relevant: fibrotic nodules similar to silicotic nodules found in humans after occupational exposure
Ability to visualize silica deposits 
Strain dependence: C3H/HeN> MRL/MpJ> NZB> Balb/c (aerosol), C57Bl/6> CBA/J (intratracheal)
Variation in fibrotic response to different formulations 
Asbestos Mouse Asbestos fibre deposition induces oxidative stress, inflammation and epithelial injury
Central injury (intratracheal) model, subpleural injury (inhalation model)
Routes of administration: i.t. (Amphibole fibres, d7–14) Inhaled (Chryostile fibres, >d28) 
Clinically relevant: features faithful to asbestosis in humans
Development of fibrotic foci is observed
Fibrosis is persistent and progressive 
Intratracheal model: uneven distribution between lobes
Inhalation model: long duration (>d28), requires specialized inhalation chambers 
Humanized models Mouse i.v. administration of human lung fibroblasts (IPF patients only, normal fibroblasts did not induce fibrosis) results in fibrotic alveolar remodelling
Fibrosis confirmed (histology and hydroxyproline)
No effect on other organs 
Fibroblast autonomous injury: permits focus on epithelial–mesenchymal interactions
Rapid onset (d30–35)
Persistence (>d63)
Opportunity to study human specific therapeutic targets 
Fibrotic foci observed in blood vessels and not lung parenchyma
NOD/SCID mice are expensive and require specialized housing
Development of fibrosis is artificial due to the absence of human immune system 
Targeted Type II cell injury Mouse Expression of the diphtheria toxin receptor under the control of the alveolar Type II cell specific promoter, surfactant protein-C
Repetitive delivery of diphtheria toxin (d14 intraperitoneal)
Selective ATII cell injury: hyperplastic ATII response (d7), fibrosis
(histology and hydroxyproline (d21–28)
(Ly)6C high monocytes and alternatively activated macrophages associated with the development of injury 
Mimics hypothesized ATII epithelial injury based aetiology of IPF 25% mortality rate at diphtheria toxin doses of 8 μg/kg; increased mortality w/o increase in collagen content at higher doses 10 μg/kg 
Spontaneous fibrosis Dog*
Donkey 
Spontaneous fibrotic lesions appear with age Analogous to human disease As yet awaiting full characterization 
    Difficult to standardize and control experiments 
ModelSpeciesFeatures/practicalitiesStrengthsWeaknesses
Bleomycin Mouse Direct cellular injury due to DNA scission
Inflammatory infiltration (d0–7 in single dose model, more persistent in multiple dose model) and vascular leak followed by fibrotic phase
Multiple delivery routes intratracheal, oropharyngeal: (Single dose model: d14–28
Multiple dose: 16 weeks
Other routes: i.v., i.t., i.p. 
Single dose model very well characterized model
Clinically relevant (bleomycin toxicity)
Rapid time frame for development and resolution (reports of resolution are conflicting)
Multiple dose model (intratracheal)
Mimics clinically relevant repetitive lung injury aetiology of PF
Prominent AEC hyperplasia, robust fibrosis similar to human pathology 
Single dose model (i.t./o.p.): Significant inflammatory involvement in early phases (neutrophils, d0–7)
Self-limiting without chronic progression (d28+, reports of resolution are conflicting)
Lack of alveolar epithelial cell hyperplasia seen in human disease
Strain specific (C57Bl6> SV> Balbc) Lot variability of bleomycin activity
Multiple dose model (i.t.): long duration of model (16 weeks), late attenuation of neutrophilia 
FITC Mouse As above Ability to visualize areas of lung injury by characteristic green fluorescence
Time frame for development of fibrosis is d14–28
Fibrotic response persists for >6 months
Not strain dependent 
Not clinically relevant
Lot variability of FITC
FITC must be made fresh and is subject to user variability 
TGFβ overexpression Rat Adenoviral delivery or doxycycline regulated transgene expression in epithelium
TGFβ expression (d1–7)
Early mononuclear infiltration (d3–7)
Alveolar consolidation, epithelial apoptosis and collagen accumulation 
Mimics human disease
Enables single growth
Factor dissection of pathology
Persistent fibrotic injury 
Early inflammatory phase due to adenoviral vector
Strain dependence C57Bl6> Balb/c 
Irradiation Mouse Free radical mediated DNA damage
Induction of inflammation followed by fibrosis
Vascular remodelling reminiscent of PAH 
Clinically relevant
Low mortality 
Long duration of model
Need to protect other organs from irradiation
Expensive
Strain dependence: C57Bl6> C3H/HeJ and CBA/J (resistant) 
Silica Mouse Silica retained in lung inducing inflammatory response and development of fibrotic nodules around deposits
Characterized by activation of the inflammasome activation
Multiple routes of administration: aerosol (d40–120), i.t./o.p. (d14–28)
Baking of silica required to destroy endotoxins 
Clinically relevant: fibrotic nodules similar to silicotic nodules found in humans after occupational exposure
Ability to visualize silica deposits 
Strain dependence: C3H/HeN> MRL/MpJ> NZB> Balb/c (aerosol), C57Bl/6> CBA/J (intratracheal)
Variation in fibrotic response to different formulations 
Asbestos Mouse Asbestos fibre deposition induces oxidative stress, inflammation and epithelial injury
Central injury (intratracheal) model, subpleural injury (inhalation model)
Routes of administration: i.t. (Amphibole fibres, d7–14) Inhaled (Chryostile fibres, >d28) 
Clinically relevant: features faithful to asbestosis in humans
Development of fibrotic foci is observed
Fibrosis is persistent and progressive 
Intratracheal model: uneven distribution between lobes
Inhalation model: long duration (>d28), requires specialized inhalation chambers 
Humanized models Mouse i.v. administration of human lung fibroblasts (IPF patients only, normal fibroblasts did not induce fibrosis) results in fibrotic alveolar remodelling
Fibrosis confirmed (histology and hydroxyproline)
No effect on other organs 
Fibroblast autonomous injury: permits focus on epithelial–mesenchymal interactions
Rapid onset (d30–35)
Persistence (>d63)
Opportunity to study human specific therapeutic targets 
Fibrotic foci observed in blood vessels and not lung parenchyma
NOD/SCID mice are expensive and require specialized housing
Development of fibrosis is artificial due to the absence of human immune system 
Targeted Type II cell injury Mouse Expression of the diphtheria toxin receptor under the control of the alveolar Type II cell specific promoter, surfactant protein-C
Repetitive delivery of diphtheria toxin (d14 intraperitoneal)
Selective ATII cell injury: hyperplastic ATII response (d7), fibrosis
(histology and hydroxyproline (d21–28)
(Ly)6C high monocytes and alternatively activated macrophages associated with the development of injury 
Mimics hypothesized ATII epithelial injury based aetiology of IPF 25% mortality rate at diphtheria toxin doses of 8 μg/kg; increased mortality w/o increase in collagen content at higher doses 10 μg/kg 
Spontaneous fibrosis Dog*
Donkey 
Spontaneous fibrotic lesions appear with age Analogous to human disease As yet awaiting full characterization 
    Difficult to standardize and control experiments 
*

West Highland Terrier.

Bleomycin is a chemotherapeutic antibiotic derived from the Actinobacterium Streptomyces verticillus, with its anti-tumour mechanism resulting from metal ion chelation and superoxide generation inducing DNA strand cleavage and interruption of the cell cycle [61]. Currently used clinically as a treatment for various cancers, pulmonary toxicity represents a major clinical side effect, manifesting as cough, dyspnoea, fever, cyanosis and deterioration of lung function parameters and progression to fibrosis in some cases. This effect predominates in the lung as a result of low-organ-specific levels of bleomycin hydrolase, resulting in impaired clearance of the active agent [46]. Indeed, variability in levels of bleomycin hydrolase across mouse strains is thought to account for the well-established strain-dependent variability observed in response to bleomycin, with C57Bl6 mice exhibiting greater sensitivity than that BALB/c to bleomycin injury [46].

Ease of administration in a range of small animal settings, rapid onset of fibrosis in addition to well established end points have undoubtedly been contributory factors to common use of the bleomycin model. When administered, as most frequently used in mice, via a single [oropharyngeal or intratracheal (i.t.)] instillation directly into the respiratory tract normally at a dose of 1–2 mg/kg (approximately equivalent to 0.025–0.05 U/mouse) [62], bleomycin induces epithelial injury and subsequent inflammation. This is marked by neutrophil infiltration, macrophage activation and up-regulation of pro-inflammatory cytokines and chemokines, accompanied by significant vascular leak, leading some to utilize the inflammatory phase of bleomycin as a useful model of acute lung injury [63]. The inflammatory phase persists for 1 week to 10 days and precedes fibrosis which typically peaks around 3–4 weeks post initial challenge. Significant permutations from this include a recently described multiple hit model involving eight bi-weekly low dose i.t. administrations of bleomycin, designed to mimic the multiple episodes of epithelial injury thought to underlie the aetiology of IPF [64]. The persistent fibrotic lesion which is induced resembles many of the classical features of UIP without excessive neutrophilic inflammation; however, it requires a significant time commitment with implications for experimental costs.

In the single-hit bleomycin model, the fibrotic phase is characterized by an expanding population of myofibroblasts, excessive ECM production and increased levels of pro-fibrotic cytokines such as TGFβ. Fibrotic end points are most commonly assessed by semi quantitative histological scoring, requiring extensive histological sampling and multiuser analysis to avoid sampling bias, or biochemical quantitation of collagen or preferably its enriched component, the amino acid hydroxyproline; effectively a specific marker for collagen which can be readily obtained from lung homogenates by acid hydrolysis and quantified by HPLC. Collagen forms an important component of the naïve lung and as such, the global increase in collagen content observed during fibrotic injury is generally only 2-fold above base line, often proving an unsatisfactorily blunt instrument for pharmacodynamic assessment of anti-fibrotics. Recent studies have investigated the use of in vivo and ex vivo imaging as a means of improving the signal window, also introducing the possibility of longitudinal assessment of the model [6569]. In particular Scotton et al. [69] have characterized the bleomycin model using high resolution ex vivo microCT of injured murine lungs, comparing this with assessment of collagen content by both sircoll assay and hydroxyproline HPLC. Use of microCT significantly increased the signal window for fibrotic injury at the 28 day peak of fibrosis, as well as providing morphological insight of the time course of the bleomycin injury at the whole tissue level; detailing the rapid progression of established fibrosis in the later stages of injury (14–28 days) and highlighting that while the fibrotic lesions did not appear to progress beyond 28 days, lung architecture remains grossly remodelled with consolidating deposition of ECM at least up to 6 months after original insult. It is clear that thorough understanding of the natural history of the bleomycin injury is critical for the design of mechanistic studies with anti-fibrotic interventions. In their seminal review, Moeller et al. [46] reflected on the apparent failure of the bleomycin model to predict translatability of anti-fibrotics, suggesting that dosing of therapeutics in the fibrotic phase of the injury (from day 10 to 14), rather than the early inflammatory phase, should be adopted in order to enhance our ability to dissect anti-inflammatory from anti-fibrotic drug profiles [46]. In support of this, the TGFβ-R1 (Alk5) inhibitor SB525334 is efficacious when dosed during the fibrotic phase of the bleomycin model [69], and use of therapeutic dosing of recent candidate anti-fibrotics, including pirfenidone [70], nintedanib [71], imatinib [72] and prednisolone [73], has arguably been predictive of clinical outcome.

In further development of lung fibrosis models, the role of multiple injuries and risk factors such as age and sex may also need to be considered in model design. Chronic viral infection has controversially been proposed as a cause of ongoing epithelial injury in IPF, and recent studies have suggested that γ-Herpes virus infection exacerbated bleomycin induced fibrosis where as other infective agents including Pseudomonas aerginosa and H1N1 failed to augment the fibrotic response [74]. Given that the incidence of IPF increases with advancing years, particularly in the male population, both the age and sex of experimental animals may also warrant important consideration in ‘ideal’ model design. In this regard, recent studies have shown that aged mice are more susceptible to bleomycin induced fibrosis than younger animals [75] with aged male mice potentially more sensitive than aged female mice [76].

In summary, the bleomycin model of pulmonary fibrosis is the most commonly used and best characterized in vivo models of lung injury and fibrosis currently at our disposal. Although it does not fully recapitulate all of the features seen in human disease, it remains a rapid, reproducible and therefore cost-effective model for interrogating key fibrotic mechanisms, when used in the correct context. This requires a thorough understanding of the model, its limitations and the specific mechanistic questions to be addressed by its usage.

TRANSLATIONAL MODELS OF ASTHMA AND COPD EXACERBATIONS

The best general definition of an exacerbation of COPD is a sustained worsening of the patient's condition, from the stable state and beyond normal day–day variations, that is acute in onset and may warrant additional treatment in patients with underlying COPD [77]. The most common cause of COPD exacerbations are believed to be bacteria and viruses, with one study detecting viral and/or bacterial infection in 78% of patients admitted to hospital for exacerbations of their disease (29.7% bacterial, 23.4% viral and 25% viral/bacterial co-infection) [78]. The most frequently identified respiratory viruses in COPD exacerbations are rhinoviruses, influenza viruses, coronaviruses, and respiratory syncytial virus (RSV) [79] and the most commonly isolated bacterial pathogens are H. influenzae and Moraxella catarrhalis. More rarely parainfluenza viruses and human metapneumoviruses (HMPV) have also been detected [78,8082]. Viral and bacterial COPD exacerbations are associated with impaired lung function (exacerbations with co-infection tend to have more marked lung function impairment) together with increases in inflammatory cells and pro-inflammatory mediators in sputum [78,83]. Increased sputum neutrophilia correlates with exacerbation severity regardless of whether the exacerbation is associated with bacteria or viral infection [78] however sputum eosinophils have proven to be a good predictor of viral-associated exacerbations [78]. Although corticosteroids can reduce duration and frequency of exacerbations, they have limited effects on sputum neutrophils.

Acute exacerbations of asthma are caused by an exaggerated lower airway response to an environmental exposure with the classic symptoms being shortness of breath, wheezing and chest tightness. Respiratory viral infection is the most common cause of a severe asthma exacerbation, but exacerbations can also be triggered by allergen, environmental pollutants, occupational sensitizers/irritants or medications (e.g. aspirin). Rhinoviruses are the most commonly detected infectious agents in asthma exacerbations. In addition to rhinoviruses, other viruses, such as RSV, influenza viruses, coronaviruses, HMPV, parainfluenza viruses, adenoviruses, and bocaviruses, have all been detected in subjects experiencing an asthma exacerbation. Exacerbations are characterized by a fall in forced expiratory volume in 1 s (FEV1) or PEF, mucus overproduction, airways inflammation and AHR. The precise nature of the inflammation depends on the specific trigger of the exacerbation, but increased neutrophils and eosinophils have been observed following rhinovirus infection [84]. Individuals who are corticosteroid insensitive suffer from frequent exacerbations of disease, which further decrease corticosteroid sensitivity [84].

Various pre-clinical models have been established and published in the scientific literature in rodents and rabbits to try and mimic characteristic features of bacterial and viral-induced exacerbations in asthma and COPD (Table 5). These model systems involve challenging animals which have been previously exposed to allergen (for asthma exacerbations) or TS (for COPD exacerbations) with either bacteria, components of the outer membrane of Gram negative bacteria (e.g. LPS), viruses or a viral mimetic (e.g. polyinosinic, PolyIC). The features, strengths and weaknesses of the models were presented at the 2nd CCRS and are summarized in Table 5 and below. All but one of the models described in Table 5 attempt to model viral-induced exacerbations of asthma by infecting animals previously sensitized to allergen (HDM or OVA) with either a live virus or the viral mimetic, PolyIC. Several of these model systems exhibited an exaggerated airway inflammatory response in the animals exposed to allergen with virus compared with animals just exposed to allergen [8587] (Bal et al., de-Alba et al. and Shaw et al. in [3]), whereas others did not [85,88] (Barrett et al. and Shaw et al. in [3]). The human rhinovirus (HRV)1B virus was used in three of the models where an exaggerated airway inflammatory phenotype was not seen. In one of these models (Shaw et al. in [3]), it was noted that the virus did not replicate well, whereas in a separate model [88], an impaired antiviral immune response in the presence of chronic allergic inflammation was observed. Several of the model systems found that viral challenge of animals previously exposed to allergen resulted in increased AHR compared with animals not exposed to virus [86] (Bal et al. and de-Alba et al. in [3]). Some of the models were corticosteroid insensitive [85], whereas others had mixed sensitivity depending on the disease feature being studies [86,87]. Finally, experimenters have sought to model a viral-induced COPD exacerbation by infecting mice that had previously been exposed to cigarette smoke, with PolyIC (Russell et al. in [3,29]) or live influenza A virus [8991]. These models were able to mimic the exaggerated airway inflammatory response observed during an exacerbation. In the study in which mice were exposed to the combination of cigarette smoke and PolyIC, the exaggerated inflammatory response was partially sensitive to corticosteroids, Roflumilast (the PDE4 inhibitor which was relatively recently licensed for the treatment of COPD) and IL-6 neutralization. In one of the studies in which mice were exposed to a combination of cigarette smoke and live influenza A virus, dexamethasone treatment partially attenuated the inflammatory response in the BAL of smoke-exposed, virally infected animals [90]. In contrast with controls, however, dexamethasone-treated smoke-exposed influenza-infected mice had a worsened health status. Interestingly, in the present study, treatment of virally infected smoke-exposed mice with the peroxisome-proliferator-activated receptor-γ (PPAR-γ) agonist, Pioglitazone, proved more efficacious than the steroid intervention.

Table 5
Exacerbation models

Summary of exacerbation models presented at the 2nd CCRS, held in Horsham, September 2012. For the purpose of the table ‘acute’ is defined as 3 weeks or less and ‘chronic’ as more than 3 weeks.

ModelSpeciesTrying to modelMethodsStrengths/featuresWeaknesses
Acute HDM + PolyIC Mouse Viral asthma exacerbation HDM (s.c.), 14 days later
HDM (i.n.) and single dose of PolyIC (i.n.) before, during or after HDM (i.n.) challenge [16]
HDM (i.n) for 3 weeks and single dose PolyIC (i.n) [67
Reproduces similar features of human viral asthma exacerbation, e.g. enhanced BAL neutrophils, lymphocytes, macrophages, increased AHR to methacholine
Increased BAL neutrophils and BAL CD8+ T-cells, increased CD4+/CD8+ activation, increased KC, IL-1β, IFN-γ compared with HDM alone
Insensitive to corticosteroids 
Viral mimetic
Less widely used and characterized as yet
Does not replicate chronicity of inflammation observed in asthma patients which can develop over years prior to an exacerbation 
Acute HDM + Influenza Mouse Viral asthma exacerbation HDM (i.n.) for 3 days, challenged 2 weeks later (HDM; i.n.) and then infected with influenza 2 days later (Bal et al. in [2]) Reproduces similar features of asthma exacerbation, e.g. increased BAL eosinophils, increased IL-5 and reduced IFN-γ in lung lysate, trend towards increase AHR to methacholine Does not replicate chronicity of inflammation observed in asthma patients which can develop over years prior to an exacerbation 
Chronic HDM + RSV Mouse Viral asthma exacerbation 5 week HDM and 1 dose of RSV [69Reproduces similar features of human viral asthma exacerbation, e.g. increased BAL neutrophils, lymphocytes, eosinophils, macrophages, exhaled NO, BAL mucin, TNF-α, KC, IL-2, IL-10, IL-1β, IL-5 and decreased IL-4 compared with HDM challenged mice
Mixed sensitivity to prednisolone (no effect on macrophages or neutrophils but inhibition of lymphocytes and eosinophils and exhaled NO)
Corticosteroids have been shown to reduce eosinophils and exhaled NO in asthma exacerbations 
RSV inhibited HDM-elicited eosinophilia
Less widely used and characterized as yet 
Chronic HDM + influenza Mouse Viral asthma exacerbation 5 week HDM and on day 35 influenza A/
HKX31[H3N2] (Barrett et al. in [1]) 
 No further increase in BAL cells with HDM/influenza combination
Efficacy of a steroid in presence of infection was reversed and lead to an exacerbation; not clear that this is consistent with clinical data 
   7 week HDM and H3N2 (i.n. single dose) (Shaw et al. in [1]) Reproduces similar features of human viral exacerbation (increased BAL eosinophils, neutrophils) Steroid responsive; no change in PO2 readout 
   5 week HDM + influenza A (1 dose) [69Reproduces similar features of human viral asthma exacerbation, e.g. increased BAL lymphocytes, neutrophils, macrophages, IL-2, IL-4, IL-10, IL-12 and reduced eosinophils and IFN-γ compared with HDM challenged mice
Mixed sensitivity to prednisolone (no effect on macrophages or neutrophils but inhibition of lymphocytes, eosinophils and exhaled NO) 
 
Chronic HDM + HRV Mouse Viral asthma exacerbation 7 week HDM and HRV1B (i.n. on last 3 days) [70]
3 week HDM and HRV1B (i.n. single dose after last HDM) [67]
7 week HDM and HRV1B (i.n. single dose) (Shaw et al. in [1]) 
Impaired antiviral immune response in presence of chronic allergic inflammation (i.e. in mice chronically exposed to HDM and subsequently exposed to HRV1B compared with mice just exposed to HRV1B), e.g. reduced BAL neutrophils, and reduced antiviral and proinflammatory BAL cytokines (e.g. IFN-α, IFN-γ, IL-12) No exacerbation of eosinophilic inflammation or AHR in mice exposed to HDM and HRV1B versus HDM alone Steroid responsive (inflammation suppressed)
No exacerbation phenotype
No exacerbation phenotype (virus did not replicate well) 
OVA and human Parainfluenza virus (hPIV) Guinea pig Viral asthma exacerbation OVA sensitized, with 2× inoculations of hPIV-3 day 10 and 11 and OVA challenged day 14 [68Reproduces some features of human viral asthma exacerbation (e.g. increased inflammatory cells in BAL, increased airways obstruction, airways hyper-responsiveness)
Exacerbated airway obstruction and AHR were insensitive to corticosteroid
Inflammation was partially inhibited by a steroid
Greater similarities with human lung anatomy and pharmacology than rodents 
Non-availability of transgenics and knockouts
Restricted availability of reagents
Does not replicate chronicity of inflammation observed in asthma patients which can develop over years prior to an exacerbation 
Acute smoke and PolyIC Mouse Viral COPD exacerbation TS for 4 days and then PolyIC (i.n)-(Russell et al. in [1]) Exaggerated inflammatory response which was significantly greater than the additive effect of the two stimuli
Enhanced inflammatory response was partially sensitive to steroids and roflumilast 
Viral mimetic
Less widely used and characterized as yet
Does not replicate chronicity of inflammation observed in COPD patients which can develop over years prior to an exacerbation 
   TS for 2 weeks with 3 doses of PolyIC (i.n.) on day 0, 3 and 7 [29Reproduces some features of human viral COPD exacerbation, e.g. increase in BAL neutrophils and lymphocytes in PolyIC/smoke exposed mice versus smoke alone
Largely corticosteroid insensitive, e.g. dex did not inhibit BAL neutrophilia or inhibit BAL cytokine/chemokines; actually increased BAL GMCSF levels, but dex did inhibit BAL lymphocytes and macrophages 
 
ModelSpeciesTrying to modelMethodsStrengths/featuresWeaknesses
Acute HDM + PolyIC Mouse Viral asthma exacerbation HDM (s.c.), 14 days later
HDM (i.n.) and single dose of PolyIC (i.n.) before, during or after HDM (i.n.) challenge [16]
HDM (i.n) for 3 weeks and single dose PolyIC (i.n) [67
Reproduces similar features of human viral asthma exacerbation, e.g. enhanced BAL neutrophils, lymphocytes, macrophages, increased AHR to methacholine
Increased BAL neutrophils and BAL CD8+ T-cells, increased CD4+/CD8+ activation, increased KC, IL-1β, IFN-γ compared with HDM alone
Insensitive to corticosteroids 
Viral mimetic
Less widely used and characterized as yet
Does not replicate chronicity of inflammation observed in asthma patients which can develop over years prior to an exacerbation 
Acute HDM + Influenza Mouse Viral asthma exacerbation HDM (i.n.) for 3 days, challenged 2 weeks later (HDM; i.n.) and then infected with influenza 2 days later (Bal et al. in [2]) Reproduces similar features of asthma exacerbation, e.g. increased BAL eosinophils, increased IL-5 and reduced IFN-γ in lung lysate, trend towards increase AHR to methacholine Does not replicate chronicity of inflammation observed in asthma patients which can develop over years prior to an exacerbation 
Chronic HDM + RSV Mouse Viral asthma exacerbation 5 week HDM and 1 dose of RSV [69Reproduces similar features of human viral asthma exacerbation, e.g. increased BAL neutrophils, lymphocytes, eosinophils, macrophages, exhaled NO, BAL mucin, TNF-α, KC, IL-2, IL-10, IL-1β, IL-5 and decreased IL-4 compared with HDM challenged mice
Mixed sensitivity to prednisolone (no effect on macrophages or neutrophils but inhibition of lymphocytes and eosinophils and exhaled NO)
Corticosteroids have been shown to reduce eosinophils and exhaled NO in asthma exacerbations 
RSV inhibited HDM-elicited eosinophilia
Less widely used and characterized as yet 
Chronic HDM + influenza Mouse Viral asthma exacerbation 5 week HDM and on day 35 influenza A/
HKX31[H3N2] (Barrett et al. in [1]) 
 No further increase in BAL cells with HDM/influenza combination
Efficacy of a steroid in presence of infection was reversed and lead to an exacerbation; not clear that this is consistent with clinical data 
   7 week HDM and H3N2 (i.n. single dose) (Shaw et al. in [1]) Reproduces similar features of human viral exacerbation (increased BAL eosinophils, neutrophils) Steroid responsive; no change in PO2 readout 
   5 week HDM + influenza A (1 dose) [69Reproduces similar features of human viral asthma exacerbation, e.g. increased BAL lymphocytes, neutrophils, macrophages, IL-2, IL-4, IL-10, IL-12 and reduced eosinophils and IFN-γ compared with HDM challenged mice
Mixed sensitivity to prednisolone (no effect on macrophages or neutrophils but inhibition of lymphocytes, eosinophils and exhaled NO) 
 
Chronic HDM + HRV Mouse Viral asthma exacerbation 7 week HDM and HRV1B (i.n. on last 3 days) [70]
3 week HDM and HRV1B (i.n. single dose after last HDM) [67]
7 week HDM and HRV1B (i.n. single dose) (Shaw et al. in [1]) 
Impaired antiviral immune response in presence of chronic allergic inflammation (i.e. in mice chronically exposed to HDM and subsequently exposed to HRV1B compared with mice just exposed to HRV1B), e.g. reduced BAL neutrophils, and reduced antiviral and proinflammatory BAL cytokines (e.g. IFN-α, IFN-γ, IL-12) No exacerbation of eosinophilic inflammation or AHR in mice exposed to HDM and HRV1B versus HDM alone Steroid responsive (inflammation suppressed)
No exacerbation phenotype
No exacerbation phenotype (virus did not replicate well) 
OVA and human Parainfluenza virus (hPIV) Guinea pig Viral asthma exacerbation OVA sensitized, with 2× inoculations of hPIV-3 day 10 and 11 and OVA challenged day 14 [68Reproduces some features of human viral asthma exacerbation (e.g. increased inflammatory cells in BAL, increased airways obstruction, airways hyper-responsiveness)
Exacerbated airway obstruction and AHR were insensitive to corticosteroid
Inflammation was partially inhibited by a steroid
Greater similarities with human lung anatomy and pharmacology than rodents 
Non-availability of transgenics and knockouts
Restricted availability of reagents
Does not replicate chronicity of inflammation observed in asthma patients which can develop over years prior to an exacerbation 
Acute smoke and PolyIC Mouse Viral COPD exacerbation TS for 4 days and then PolyIC (i.n)-(Russell et al. in [1]) Exaggerated inflammatory response which was significantly greater than the additive effect of the two stimuli
Enhanced inflammatory response was partially sensitive to steroids and roflumilast 
Viral mimetic
Less widely used and characterized as yet
Does not replicate chronicity of inflammation observed in COPD patients which can develop over years prior to an exacerbation 
   TS for 2 weeks with 3 doses of PolyIC (i.n.) on day 0, 3 and 7 [29Reproduces some features of human viral COPD exacerbation, e.g. increase in BAL neutrophils and lymphocytes in PolyIC/smoke exposed mice versus smoke alone
Largely corticosteroid insensitive, e.g. dex did not inhibit BAL neutrophilia or inhibit BAL cytokine/chemokines; actually increased BAL GMCSF levels, but dex did inhibit BAL lymphocytes and macrophages 
 

Thus, from the published and unpublished (discussed at the CCRS) experiences, the HRV models have not been easy to transfer between laboratories, using minor group RV in wild-type mice or major group RV in intercellular adhesion molecule-1 (ICAM-1) transgenics. On the other hand flu or RSV may be more robust/reproducible live virus exacerbation agents from the experiences of reproducing these in different laboratories. Viral mimics such as PolyIC are likely to be useful, but do not reproduce all the effects seen with replicating live viruses. On the other hand, the live virus models have significant challenges in achieving the right dose and strain of virus to get a reproducible but well-tolerated response, as well as requiring special containment for safe handling.

TRANSLATIONAL IN VITRO AND EXPLANT MODELS OF RESPIRATORY DISEASE

The issues of species differences (Table 1) and the mixed record of clinical predictivity of animal models of respiratory diseases (Tables 25) have led to an increased focus on the development and use of in vitro and explant models, using human cells and tissues (Figure 1 and Table 6). Using tissue or cells from the human airways or lungs has advantages in that this reflects local changes/differences that occur at the site of disease in humans suffering from that disease. Examples are given in Table 6, along with some of the pros and cons of these systems. Moreover, analogous studies with cells and tissues from animal models can also be carried out in vitro, in order to understand the translation from these in vivo models through in vitro systems to human clinical diseases.

Table 6
Human in vitro and explant models of lung diseases
ModelCell typesStrengths/featuresWeaknesses
Lung slices Parenchymal cells
Resident leucocytes* 
Studies the appropriate target tissue
Enables study of the pharmacology of small airways and lung blood vessels as well as mediator release
Can be challenged by a range of agents, e.g. allergen (following sensitization), LPS, TS extract 
Lacks the integrated system in vivo, especially neuronal, endocrine and circulatory elements
Not usually feasible to secure disease lungs, so in vitro conditioning to mimic the disease required (e.g. sensitization and challenge to study asthmatic responses
Logistics and cost of sourcing the tissue (surgical or post-mortem) 
Lung fragments Parenchymal cells Studies the appropriate target tissue Lacks the integrated system in vivo, especially neuronal, endocrine and circulatory elements 
 Resident cells* Enables study mediator release from lung tissue Can be challenged by a range of agents, e.g. allergen (following sensitization), LPS Not usually feasible to secure disease lungs, so in vitro conditioning to mimic the disease required (e.g. sensitization and challenge to study asthmatic responses 
   Not feasible to study airway or blood vessel pharmacology 
   Logistics and cost of sourcing the tissue (surgical or post-mortem) 
Biopsy explants Parenchymal cells
Resident/
migrated leucocytes* 
Studies the appropriate tissue, potentially from the disease of concern
Enables study of mediator release from disease and control lung tissue
May include inflammatory cell types present in the disease
Potential to study basal as well as stimulated mediator release 
Lacks the integrated system in vivo, especially neuronal, endocrine and circulatory elements.
Safety, consent, logistics and cost of sourcing the biopsies (via bronchoscopy) 
Air–liquid interface cultures Lung epithelial cells Studies a key cell type from the appropriate target tissue
Enable study of mediator release from ‘normal’ and diseased cells Potential to study basal as well as stimulated mediator release
Enables measurement of mucus secretion, ciliary beat and mucociliary clearance
Potential to study basal as well as stimulated mediator release 
Lacks the integrated system in vivo, especially neuronal, endocrine and circulatory elements.
Very time consuming and labour intensive; ~4 weeks to air–liquid interface; require regular feeding of cultures, contamination, cost 
  Enables measurement of ion and fluid transport  
  Enables measurement of bacterial and viral infection Enables study of systemic drug absorption of inhaled drugs  
Bronchospheres Lung epithelial cells Studies a key cell type from the appropriate target tissue Not yet well characterized or widely used 
  In vitro model system of respiratory tract glandular acini and hyperplasia  
Cellular co-cultures Selection Allows the study of cell–cell interactions
Can combine two or more key cell types 
Difficult to know which combinations to study
Difficult to be comprehensive in covering all the lung cell types with a native phenotype 
Peripheral blood mononuclear cells Circulating leucocytes Enables study of mediator release
Potential to study basal as well as stimulated mediator release
Enables study of mediator release from normal and diseased cells
Enables study of proliferation
Enables study of cell migration
Enables study of compound cytotoxicity 
Lacks the integrated cellular systems in vivo
Few mast cells; basophils do not have the same biology
Lung parenchymal cells (epithelial, smooth muscle) not represented 
Isolated neurons Neuronal Studies neurons, which play roles in human lung diseases Difficult to isolate and maintain appropriate neurons to represent those involved in lung biology 
ModelCell typesStrengths/featuresWeaknesses
Lung slices Parenchymal cells
Resident leucocytes* 
Studies the appropriate target tissue
Enables study of the pharmacology of small airways and lung blood vessels as well as mediator release
Can be challenged by a range of agents, e.g. allergen (following sensitization), LPS, TS extract 
Lacks the integrated system in vivo, especially neuronal, endocrine and circulatory elements
Not usually feasible to secure disease lungs, so in vitro conditioning to mimic the disease required (e.g. sensitization and challenge to study asthmatic responses
Logistics and cost of sourcing the tissue (surgical or post-mortem) 
Lung fragments Parenchymal cells Studies the appropriate target tissue Lacks the integrated system in vivo, especially neuronal, endocrine and circulatory elements 
 Resident cells* Enables study mediator release from lung tissue Can be challenged by a range of agents, e.g. allergen (following sensitization), LPS Not usually feasible to secure disease lungs, so in vitro conditioning to mimic the disease required (e.g. sensitization and challenge to study asthmatic responses 
   Not feasible to study airway or blood vessel pharmacology 
   Logistics and cost of sourcing the tissue (surgical or post-mortem) 
Biopsy explants Parenchymal cells
Resident/
migrated leucocytes* 
Studies the appropriate tissue, potentially from the disease of concern
Enables study of mediator release from disease and control lung tissue
May include inflammatory cell types present in the disease
Potential to study basal as well as stimulated mediator release 
Lacks the integrated system in vivo, especially neuronal, endocrine and circulatory elements.
Safety, consent, logistics and cost of sourcing the biopsies (via bronchoscopy) 
Air–liquid interface cultures Lung epithelial cells Studies a key cell type from the appropriate target tissue
Enable study of mediator release from ‘normal’ and diseased cells Potential to study basal as well as stimulated mediator release
Enables measurement of mucus secretion, ciliary beat and mucociliary clearance
Potential to study basal as well as stimulated mediator release 
Lacks the integrated system in vivo, especially neuronal, endocrine and circulatory elements.
Very time consuming and labour intensive; ~4 weeks to air–liquid interface; require regular feeding of cultures, contamination, cost 
  Enables measurement of ion and fluid transport  
  Enables measurement of bacterial and viral infection Enables study of systemic drug absorption of inhaled drugs  
Bronchospheres Lung epithelial cells Studies a key cell type from the appropriate target tissue Not yet well characterized or widely used 
  In vitro model system of respiratory tract glandular acini and hyperplasia  
Cellular co-cultures Selection Allows the study of cell–cell interactions
Can combine two or more key cell types 
Difficult to know which combinations to study
Difficult to be comprehensive in covering all the lung cell types with a native phenotype 
Peripheral blood mononuclear cells Circulating leucocytes Enables study of mediator release
Potential to study basal as well as stimulated mediator release
Enables study of mediator release from normal and diseased cells
Enables study of proliferation
Enables study of cell migration
Enables study of compound cytotoxicity 
Lacks the integrated cellular systems in vivo
Few mast cells; basophils do not have the same biology
Lung parenchymal cells (epithelial, smooth muscle) not represented 
Isolated neurons Neuronal Studies neurons, which play roles in human lung diseases Difficult to isolate and maintain appropriate neurons to represent those involved in lung biology 
*

Leucocytes may be subject to a degree of washout in these preparations.

However, work with human materials is challenging, in terms of logistics, costs and ethical constraints; for example, it is difficult to reliably get fresh human asthmatic lung for the routine preparation of lung slices. In addition, the difficulty in obtaining fresh human primary cells and tissues, from bronchoscopy or post-mortem, and the limited quantities obtained, has led to the use of a variety of cell models and the use of blood cells as surrogates for studying cellular activation and drug efficacy.

A number of different primary cells have been isolated by positive or negative selection from explanted lung tissue, tissue resection, BAL fluid and bronchial biopsy to obtain a pure cultures of individual cell types. These include epithelial cells, airway smooth muscle cells, fibroblasts, endothelial cells, BAL macrophages and leucocytes and alveolar type II epithelial cells. Analysis of primary cells should from patients airways/lungs should provide greater understanding of the cellular and molecular processes involved in disease pathogenesis [9294].

However, due to the limited availability of these primary cells, researchers often examine mechanisms and pathways in cultured immortalized cells or monocyte-derived macrophages for example before confirming data in smaller numbers of primary cells from patients [9294]. These cell lines may retain many of the phenotypic characteristics of the primary cell and are easy to use but the use of these immortalized cells to investigate certain aspects of disease requires better validation particularly in relation to the lack of the natural tissue milieu and correct cell–cell/cell–matrix interactions [9294].

As a result, investigators have begun to use co-culture systems such as epithelial cells grown at air–liquid interface incubated with T-cells or smooth muscle and epithelial cell co-culture systems [92]. The use of epithelial cells grown in 3D culture to form airway-like tubes has been another recent development. An alternative approach is to use whole biopsies as the complex model in order to examine pathological mechanisms and drug effects [92,93,95]. The use of precision cut lung slices (PCLS) has provided extremely useful information regarding airway relaxation in response to drugs [96] and also enables the later isolation of individual cell types for multi-omics analysis [97,98]. These systems have been studied and compared at baseline and after exposure to triggers important in a variety of respiratory diseases: (i) COPD: cigarette smoke, oxidants, environmental pollutants and fine particles, cytokines [92,98]; (ii) asthma: passive sensitization and allergen challenge, complex allergens such as HDM extracts, anti-IgE cross-linking antibodies, cytokines and mixtures thereof implicated in asthma (e.g. IL-4, -5 and -13) [94]; (iii) fibrosis: incubation with growth factors (e.g. TGFβ) and cytokines implicated in fibrosis, and direct exposure to bleomycin (reviewed in [99]); and (iv) exacerbations: combinations of the above with viruses or other microbes or products such as PolyIC or LPS which mimic these, or host responses to them such as interferons.

In these experiments, differences in response of cells in culture from patients and controls are compared and related to differences in pathophysiological processes seen in tissue from the airways or lung. It is important that these cells are cultured and sub-cultured (passaged) for the minimum time necessary to prevent changes in their functional characteristics [92,94]. Explant cell/tissue models could in principle provide data more closely aligned with that seen in human disease than animal models. However, cell culture models by their very nature cannot recapitulate the full complexities of human disease pathophysiology. Primary cells from mice are often isolated and exposed to the same stimuli as human primary cells to determine whether responses and intracellular pathway activation are the same across species in order to further validate the animal model used.

There are often significant issues and uncertainties associated with the in vitro triggers/stimuli used, with variation resulting from methodological or batch differences as with cigarette smoke extract (CSE), allergen or inflammatory mediator stimulation of cells [92,94]. A major challenge for in vitro and explant systems is that it can be difficult to plausibly model some of the disease end points, e.g. emphysema in COPD, and susceptibility to exacerbations in general [92,94].

In summary, respiratory diseases are multi-faceted, with a number of pathological events which can be modelled in vitro to determine possible causative mechanisms. For example, CSE exposure and other non-smoking models have provided a wealth of data regarding the functional effects on primary cells and the activation status of pathways that have been implicated in COPD and asthma pathophysiology. A number of drugs including ICS and long acting β2-adrenergic receptor agonists (LABAs) as well as agents not yet tested in patients have been screened and selected through these models (e.g. [100]) and which will enable examination as to whether it is possible to attenuate or reverse the changes observed in respiratory disease. The translation of the data from these models into new, clinically effective medicines will determine whether these primary cell models are valid or not.

MODELLING INHALED DRUG DELIVERY

Translational modelling of inhaled agents to treat respiratory disease has some particular challenges: (i) a programme may need to choose which duration mechanism to focus on during lead discovery and optimization, and tailor the model systems used accordingly; (ii) multiple model systems are likely to be required to understand their likely duration of action in the lung depending on these mechanisms; (iii) the models need to be tailored to the specific drug discovery task and target; (iv) need to include clinical standards/comparators (including gold standards) wherever possible; (v) inhaled dose predictions from mouse to human are especially problematic; and (vi) need to factor in dose (concentration)–duration relationships.

The trend away from animal models or towards greater use of human cell and tissue systems is likely to continue: the human PCLS system is valuable, as potentially are air interface epithelial cell culture systems which allow for the apical delivery of potential inhaled medicines (Table 6). Cells and tissues from humans with respiratory disease would be valuable if possible to access, but this is in practice very difficult to achieve on a regular, reasonably frequent basis and so their use remains limited. However, dose predictions from human in vitro models seem a distant prospect, and here in particular animals can still give significant insight. Dose predictions have some use based on inhaled doses needed to achieve a pharmacological effect in dogs, NHPs, guinea pigs and rats, but probably not mice. For example, the inhaled (i.t. or intranasal using volumes that deliver drug to the lung) dose of the classic ICS fluticasone propionate, the active principle of Flovent/Flixotide, required to achieve reliable anti-inflammatory effects in mouse allergen models is of the order of 10–30 μg to a 30 g mouse (300–1000 μg/kg, [1,87,88,101,102]), whereas in mild-to-moderate asthma 100 μg is effective in a 75 kg human (~1 μg/kg).

The different mechanisms by which duration of action is achieved with inhaled drugs for chronic respiratory diseases include: (i) persistence of active drug at site of action (slow dissolution from the airway, slow clearance from the active compartment or re-uptake from blood into the lung/airway); (ii) tight binding/slow off-rate (tight binding to the target protein receptor/active site with a slow off-rate, tight binding to a target protein exosite with a slow off-rate, tight binding to a target-associated protein with a slow off-rate); or (iii) choice of a target approach with an intrinsically long duration of effect (persistent consequences of an agonist or desensitization/tolerance/receptor internalization).

SHOULD ANIMAL MODELS BE RUN LIKE RANDOMIZED CLINICAL TRIALS?

The failure to translate the myriad number of ‘successful’ studies in various animal models of respiratory diseases into ‘successful’ clinical trials has provoked widespread discussion as to the utility of animal models. Part of the problem may reside in the fact that the current respiratory disease models fail to recapitulate all facets of human disease or that specific models may reflect a single phenotype or aspect of disease that is not being tested in carefully chosen/phenotyped patients. This problem does not relate only to respiratory diseases but is common to most therapeutic areas. An issue that has increasing merit is the failure to conduct aspects of animal studies to the same rigorous standards as are necessary for human randomized controlled trials (RCTs) [103].

The introduction of the CONSORT (Consolidated Standards of Reporting Trials) statement in 1996 has led to significant improvements in the quality of all aspects of RCTs [104]. Although the fundamental requirements for generating reliable and unbiased data are very similar between RCTs and animal studies, there are key differences in the goals of the studies: defining mechanisms versus demonstrating clinical efficacy. Some aspects of RCTs may require further input into the recent ARRIVE (Animal Research: Reporting In vivo Experiments) guidelines for conduct and scientific reporting of animal studies [103,105]. These will include: (i) the blinding of the investigators to the treatment assignment until the experiment is complete in order not to bias data collection or analysis, (ii) the cause for attrition or loss of animals during the study period, (iii) calculation of sample size and methods for any subgroup analysis before the start of the experiment, and (iv) characterization of the precise disease phenotype in each animal model prior to starting the experiments [106].

The approach often taken of adding batches of animals until ‘statistical significance’ appears, is highly biased and goes against the concepts of random allocation and blinding in these studies. It is also important that animal researchers should report confidence intervals in addition to (or instead of) P values. The ARRIVE guidelines advise reporting numbers of animals and reasons for exclusion at baseline as well as comment regarding missing animals or data at outcome assessment. This may be particularly important in relation to off-target side effects [107]. It is hoped that the introduction of the ARRIVE along with the addition steps suggested above will provide a similar step-change in the quality and reporting of animal studies and a renewed trust in the utility of animal models of disease [103].

CONCLUSIONS AND FUTURE DIRECTIONS

The 2012 CCRS exchanged information on the many difficult challenges of establishing predictive animal models of respiratory diseases, that are obstacles to discovering and developing new medicines for patients suffering from these diseases, and the present article has reviewed these, focusing on asthma, COPD, fibrosis and exacerbations of respiratory disease. Of course many of these challenges apply equally to modelling other human diseases.

The respiratory research and development community is responding to these challenges in several ways: (i) greater collaboration and open sharing of data (negative and positive) between pharmaceutical companies, biotech and academia, e.g. large consortia funded by the EU/NIH/MRC together with companies in a pre-competitive manner; (ii) careful selection of the species, complexity and chronicity of the models based on a thorough and constantly evolving understanding of the biology involved; (iii) improved practices in pre-clinical research; (iv) continued refinement in models of respiratory diseases and their sub-types; (v) greater understanding of the biology underlying human respiratory diseases and their sub-types; and (vi) greater use of human (and especially disease-relevant) cells, tissues and explants.

The hope is that this will result in the discovery and development of a range of new treatments for the respiratory diseases that continue to cause substantial morbidity and mortality across the world.

Abbreviations

     
  • AHR

    airway hyper-responsiveness

  •  
  • ARRIVE

    Animal Research: Reporting In vivo Experiments

  •  
  • ATII

    alveolar type II cells

  •  
  • BAL

    bronchoalveolar lavage

  •  
  • CCRS

    Cross Company Respiratory Symposium

  •  
  • COPD

    chronic obstructive pulmonary disease

  •  
  • SCE

    cigarette smoke extract

  •  
  • ECM

    extracellular matrix

  •  
  • FCA

    Freund’s complete adjuvant

  •  
  • FEV1

    forced expiratory volume in 1 s

  •  
  • GOLD

    global initiative for COPD

  •  
  • HDM

    house dust mite

  •  
  • HRV

    human rhinovirus

  •  
  • HMPV

    human metapneumoviruses

  •  
  • ICAM-1

    intercellular adhesion molecule 1

  •  
  • ICS

    inhaled corticosteroid

  •  
  • IL

    interleukin

  •  
  • i.p.

    intraperitoneal

  •  
  • i.v.

    intravenous

  •  
  • IPF

    idiopathic pulmonary fibrosis

  •  
  • i.t.

    intratracheal

  •  
  • LAR

    late asthmatic response, LPS, lipopolysaccharide

  •  
  • MRC

    Medical Research Council

  •  
  • NCE

    new chemical entity

  •  
  • NHP

    non-human primate

  •  
  • NIH

    National Institutes of Health

  •  
  • NOD

    non-obese diabetic

  •  
  • o.p.

    oropharyngeal

  •  
  • PCLS

    precision cut lung slices

  •  
  • PDE4

    phosphodiesterase type 4

  •  
  • PEF

    peak expiratory flow

  •  
  • PolyIC

    polyinosinic: polycytidylic acid

  •  
  • RCT

    randomized controlled trial

  •  
  • RSV

    respiratory syncytial virus

  •  
  • SCID

    severe combined immunodeficiency

  •  
  • TGFβ

    transforming growth factor β

  •  
  • TGFβR1

    transforming growth factor receptor 1

  •  
  • Th

    T-helper

  •  
  • TNF

    tumour necrosis factor

  •  
  • TS

    tobacco smoke

  •  
  • U-BIOPRED

    Unbiased BIOmarkers in PREDiction of respiratory disease outcomes project

  •  
  • VEGF

    vascular endothelial growth factor

We acknowledge all who participated in the symposium and the sponsoring companies, but particularly Professor Stephen Holgate (University of Southampton) and Dr Anthony Holmes (NC3Rs) for their support and advice, and the plenary speakers Professor Stephen Holgate, Professor Dave Singh, Professor Sebastian Johnston and Dr Paul Mercer. R.G.K. and I.M.A. would also like to acknowledge the substantial contributions made by members of the U-BIOPRED Innovative Medicines Initiative project on severe asthma.

References

References
1
Holmes
 
A.M.
Solari
 
R.
Holgate
 
S.T.
 
Animal models of asthma: value, limitations and opportunities for alternative approaches
Drug Discov. Today
2011
, vol. 
16
 (pg. 
659
-
670
)
[PubMed]
2
Mullane
 
K.
Williams
 
M.
 
Animal models of asthma: reprise or reboot?
Biochem. Pharmacol.
2013
, vol. 
87
 (pg. 
131
-
139
)
3
Abbott-Banner
 
K.H.
Holmes
 
A.
Adcock
 
I.
Rao
 
N.L.
Barrett
 
E.
Knowles
 
R.
 
Models of respiratory disease symposium
J. Inflamm.
2013
, vol. 
10
 
Suppl. 1
pg. 
I1
 
4
Kumar
 
R.K.
Foster
 
P.S.
 
Are mouse models of asthma appropriate for investigating the pathogenesis of airway hyper-responsiveness?
Front. Physiol.
2012
, vol. 
3
 pg. 
312
 
[PubMed]
5
Nials
 
A.T.
Uddin
 
S.
 
Mouse models of allergic asthma: acute and chronic allergen challenge
Dis. Model. Mech.
2008
, vol. 
1
 (pg. 
213
-
220
)
[PubMed]
6
Stevenson
 
C.S.
Birrell
 
M.A.
 
Moving towards a new generation of animal models for asthma and COPD with improved clinical relevance
Pharmacol. Ther.
2011
, vol. 
130
 (pg. 
93
-
105
)
[PubMed]
7
Alderton
 
W.K.
Angell
 
A.D.
Craig
 
C.
Dawson
 
J.
Garvey
 
E.
Moncada
 
S.
Monkhouse
 
J.
Rees
 
D.
Russell
 
L.J.
Russell
 
R.J.
Schwartz
 
S.
Waslidge
 
N.
Knowles
 
R.G.
 
GW274150 and GW273629 are potent and highly selective inhibitors of inducible nitric oxide synthase in vitro and in vivo
Br. J. Pharmacol.
2005
, vol. 
145
 (pg. 
301
-
312
)
[PubMed]
8
Singh
 
D.
Richards
 
D.
Knowles
 
R.G.
Schwartz
 
S.
Woodcock
 
A.
Langley
 
S.
O’Connor
 
B.J.
 
Selective inducible nitric oxide synthase inhibition has no effect on allergen challenge in asthma
Am. J. Respir. Crit. Care Med.
2007
, vol. 
176
 (pg. 
988
-
993
)
[PubMed]
9
Evans
 
R.L.
Nials
 
A.T.
Knowles
 
R.G.
Kidd
 
E.J.
Ford
 
W.R.
Broadley
 
K.J.
 
A comparison of antiasthma drugs between acute and chronic ovalbumin-challenged guinea-pig models of asthma
Pulm. Pharmacol. Ther.
2012
, vol. 
25
 (pg. 
453
-
464
)
[PubMed]
10
Allen
 
I.C.
 
Induction of allergic airway disease using house dust mite allergen
Methods Mol. Biol.
2013
, vol. 
1032
 (pg. 
159
-
172
)
[PubMed]
11
Changani
 
K.
Pereira
 
C.
Young
 
S.
Shaw
 
R.
Campbell
 
S.P.
Pindoria
 
K.
Jordan
 
S.
Wiley
 
K.
Bolton
 
S.
Nials
 
T.
, et al 
Longitudinal characterization of a model of chronic allergic lung inflammation in mice using imaging, functional and immunological methods
Clin. Sci.
2013
, vol. 
125
 (pg. 
555
-
564
)
[PubMed]
12
Liu
 
Y.
Zhang
 
S.
Li
 
D.W.
Jiang
 
S.J.
 
Efficacy of anti-interleukin-5 therapy with mepolizumab in patients with asthma: a meta-analysis of randomized placebo-controlled trials
PLoS One
2013
, vol. 
8
 pg. 
e59872
 
[PubMed]
13
Henry
 
P.J.
Goldie
 
R.G.
 
Beta 1-adrenoceptors mediate smooth muscle relaxation in mouse isolated trachea
Br. J. Pharmacol.
1990
, vol. 
99
 (pg. 
131
-
135
)
[PubMed]
14
Riley
 
J.P.
Fuchs
 
B.
Sjoberg
 
L.
Nilsson
 
G.P.
Karlsson
 
L.
Dahlen
 
S.E.
Rao
 
N.L.
Adner
 
M.
 
MasT-cell mediators cause early allergic bronchoconstriction in guinea-pigs in vivo: a model of relevance to asthma
Clin. Sci.
2013
, vol. 
125
 (pg. 
533
-
542
)
[PubMed]
15
Curths
 
C.
Wichmann
 
J.
Dunker
 
S.
Windt
 
H.
Hoymann
 
H.G.
Lauenstein
 
H.D.
Hohlfeld
 
J.
Becker
 
T.
Kaup
 
F.J.
Braun
 
A.
Knauf
 
S.
 
Airway hyper-responsiveness in lipopolysaccharide-challenged common marmosets (Callithrix jacchus)
Clin. Sci.
2014
, vol. 
126
 (pg. 
155
-
162
)
[PubMed]
16
De Alba
 
J.
Otal
 
R.
Calama
 
E.
Gil
 
F.
Gozzard
 
N.
Miralpeix
 
M.
 
Poly I: C causes exacerbation in a murine allergic inflammation model driven by house dust mite in Freund's complete adjuvant
J. Inflamm.
2013
, vol. 
10
 
Suppl. 1
pg. 
P21
 
17
Churg
 
A.
Sin
 
D.D.
Wright
 
J.L.
 
Everything prevents emphysema: are animal models of cigarette smoke-induced chronic obstructive pulmonary disease any use?
Am. J. Respir. Cell Mol. Biol.
2011
, vol. 
45
 (pg. 
1111
-
1115
)
18
Fox
 
J.C.
Fitzgerald
 
M.F.
 
The role of animal models in the pharmacological evaluation of emerging anti-inflammatory agents for the treatment of COPD
Curr. Opin. Pharmacol.
2009
, vol. 
9
 (pg. 
231
-
242
)
[PubMed]
19
Goldklang
 
M.P.
Marks
 
S.M.
D’Armiento
 
J.M.
 
Second hand smoke and COPD: lessons from animal studies
Front. Physiol.
2013
, vol. 
4
 pg. 
30
 
[PubMed]
20
Groneberg
 
D.A.
Chung
 
K.F.
 
Models of chronic obstructive pulmonary disease
Respir. Res.
2004
, vol. 
5
 pg. 
18
 
[PubMed]
21
John
 
G.
Kohse
 
K.
Orasche
 
J.
Reda
 
A.
Schnelle-Kreis
 
J.
Zimmermann
 
R.
Schmid
 
O.
Eickelberg
 
O.
Yildirim
 
A.O.
 
The composition of cigarette smoke determines inflammatory cell recruitment to the lung in COPD mouse models
Clin. Sci.
2014
, vol. 
126
 (pg. 
207
-
221
)
[PubMed]
22
McGonigle
 
P.
Ruggeri
 
B.
 
Animal models of human disease: challenges in enabling translation
Biochem. Pharmacol.
2014
, vol. 
87
 (pg. 
162
-
171
)
[PubMed]
23
Vlahos
 
R.
Bozinovski
 
S.
 
Recent advances in pre-clinical mouse models of COPD
Clin. Sci.
2014
, vol. 
126
 (pg. 
253
-
265
)
[PubMed]
24
Wright
 
J.L.
Cosio
 
M.
Churg
 
A.
 
Animal models of chronic obstructive pulmonary disease
Am. J. Physiol. Lung Cell Mol. Physiol.
2008
, vol. 
295
 (pg. 
L1
-
L15
)
[PubMed]
25
Wright
 
J.L.
Churg
 
A.
 
Animal models of COPD: Barriers, successes, and challenges
Pulm. Pharmacol. Ther.
2008
, vol. 
21
 (pg. 
696
-
698
)
[PubMed]
26
Nikota
 
J.K.
Stampfli
 
M.R.
 
Cigarette smoke-induced inflammation and respiratory host defense: Insights from animal models
Pulm. Pharmacol. Ther.
2012
, vol. 
25
 (pg. 
257
-
262
)
[PubMed]
27
Baron
 
R.M.
Choi
 
A.J.
Owen
 
C.A.
Choi
 
A.M.
 
Genetically manipulated mouse models of lung disease: potential and pitfalls
Am. J. Physiol. Lung Cell Mol. Physiol.
2012
, vol. 
302
 (pg. 
L485
-
L497
)
[PubMed]
28
Churg
 
A.
Wright
 
J.L.
 
Testing drugs in animal models of cigarette smoke-induced chronic obstructive pulmonary disease
Proc. Am. Thorac. Soc.
2009
, vol. 
6
 (pg. 
550
-
552
)
[PubMed]
29
Hubeau
 
C.
Kubera
 
J.E.
Masek-Hammerman
 
K.
Williams
 
C.M.
 
Interleukin-6 neutralization alleviates pulmonary inflammation in mice exposed to cigarette smoke and poly (I: C)
Clin. Sci.
2013
, vol. 
125
 (pg. 
483
-
493
)
[PubMed]
30
Canning
 
B.J.
Chou
 
Y.
 
Using guinea pigs in studies relevant to asthma and COPD
Pulm. Pharmacol. Ther.
2008
, vol. 
21
 (pg. 
702
-
720
)
[PubMed]
31
Beckett
 
E.L.
Stevens
 
R.L.
Jarnicki
 
A.G.
Kim
 
R.Y.
Hanish
 
I.
Hansbro
 
N.G.
Deane
 
A.
Keely
 
S.
Horvat
 
J.C.
Yang
 
M.
, et al 
A new short-term mouse model of chronic obstructive pulmonary disease identifies a role for masT-cell tryptase in pathogenesis
J. Allergy Clin. Immunol.
2013
, vol. 
131
 (pg. 
752
-
762
)
[PubMed]
32
Li
 
F.
Wiegman
 
C.
Seiffert
 
J.M.
Zhu
 
J.
Clarke
 
C.
Chang
 
Y.
Bhavsar
 
P.
Adcock
 
I.
Zhang
 
J.
Zhou
 
X.
Chung
 
K.F.
 
Effects of N-acetylcysteine in ozone-induced chronic obstructive pulmonary disease model
PLoS One
2013
, vol. 
8
 pg. 
e80782
 
[PubMed]
33
Wiegman
 
C.H.
Li
 
F.
Clarke
 
C.J.
Jazrawi
 
E.
Kirkham
 
P.
Barnes
 
P.J.
Adcock
 
I.M.
Chung
 
K.F.
 
A comprehensive analysis of oxidative stress in the ozone-induced lung inflammation mouse model
Clin. Sci.
2014
, vol. 
126
 (pg. 
425
-
440
)
[PubMed]
34
Antunes
 
M.A.
Rocco
 
P.R.
 
Elastase-induced pulmonary emphysema: insights from experimental models
An. Acad. Bras. Cienc.
2011
, vol. 
83
 (pg. 
1385
-
1396
)
[PubMed]
35
Van der Velden
 
J.
Snibson
 
K.J.
 
Airway disease: the use of large animal models for drug discovery
Pulm. Pharmacol. Ther.
2011
, vol. 
24
 (pg. 
525
-
532
)
[PubMed]
36
Abraham
 
W.M.
 
Modeling of asthma, COPD and cystic fibrosis in sheep
Pulm. Pharmacol. Ther.
2008
, vol. 
21
 (pg. 
743
-
754
)
[PubMed]
37
Chapman
 
R.W.
 
Canine models of asthma and COPD
Pulm. Pharmacol. Ther.
2008
, vol. 
21
 (pg. 
731
-
742
)
[PubMed]
38
Ando
 
K.
Yanagita
 
T.
 
Cigarette smoking in rhesus monkeys
Psychopharmacology (Berl)
1981
, vol. 
72
 (pg. 
117
-
127
)
[PubMed]
39
Seok
 
J.
Warren
 
H.S.
Cuenca
 
A.G.
Mindrinos
 
M.N.
Baker
 
H.V.
Xu
 
W.
Richards
 
D.R.
McDonald-Smith
 
G.P.
Gao
 
H.
Hennessy
 
L.
, et al 
Genomic responses in mouse models poorly mimic human inflammatory diseases
Proc. Natl. Acad. Sci. U.S.A.
2013
, vol. 
110
 (pg. 
3507
-
3512
)
[PubMed]
40
Stevenson
 
C.S.
Sridhar
 
S.
Phillips
 
J.E.
 
Predicting drug efficacy using integrative models for chronic respiratory diseases
Inflamm. Allergy Drug Targets.
2013
, vol. 
12
 (pg. 
124
-
131
)
[PubMed]
41
Datta
 
A.
Scotton
 
C.J.
Chambers
 
R.C.
 
Novel therapeutic approaches for pulmonary fibrosis
Br. J. Pharmacol.
2011
, vol. 
163
 (pg. 
141
-
172
)
[PubMed]
42
Selman
 
M.
Pardo
 
A.
 
Idiopathic pulmonary fibrosis: an epithelial/fibroblastic cross-talk disorder
Respir. Res.
2002
, vol. 
3
 pg. 
3
 
43
Selman
 
M.
Pardo
 
A.
Kaminski
 
N.
 
Idiopathic pulmonary fibrosis: aberrant recapitulation of developmental programs?
PLoS Med.
2008
, vol. 
5
 pg. 
e62
 
[PubMed]
44
Scotton
 
C.J.
Chambers
 
R.C.
 
Molecular targets in pulmonary fibrosis: the myofibroblast in focus
Chest
2007
, vol. 
132
 (pg. 
1311
-
1321
)
[PubMed]
45
Klingberg
 
F.
Hinz
 
B.
White
 
E.S.
 
The myofibroblast matrix: implications for tissue repair and fibrosis
J. Pathol.
2013
, vol. 
229
 (pg. 
298
-
309
)
[PubMed]
46
Moeller
 
A.
Ask
 
K.
Warburton
 
D.
Gauldie
 
J.
Kolb
 
M.
 
The bleomycin animal model: a useful tool to investigate treatment options for idiopathic pulmonary fibrosis?
Int. J. Biochem. Cell Biol.
2008
, vol. 
40
 (pg. 
362
-
382
)
[PubMed]
47
Moore
 
B.
Lawson
 
W.E.
Oury
 
T.D.
Sisson
 
T.H.
Raghavendran
 
K.
Hogaboam
 
C.M.
 
Animal models of fibrotic lung disease
Am. J. Respir. Cell Mol. Biol.
2013
, vol. 
49
 (pg. 
167
-
179
)
[PubMed]
48
Moore
 
B.B.
Hogaboam
 
C.M.
 
Murine models of pulmonary fibrosis
Am. J. Physiol. Lung Cell Mol. Physiol.
2008
, vol. 
294
 (pg. 
L152
-
L160
)
[PubMed]
49
Sime
 
P.J.
Xing
 
Z.
Graham
 
F.L.
Csaky
 
K.G.
Gauldie
 
J.
 
Adenovector-mediated gene transfer of active transforming growth factor-beta1 induces prolonged severe fibrosis in rat lung
J. Clin. Invest.
1997
, vol. 
100
 (pg. 
768
-
776
)
[PubMed]
50
Kolb
 
M.
Margetts
 
P.J.
Anthony
 
D.C.
Pitossi
 
F.
Gauldie
 
J.
 
Transient expression of IL-1beta induces acute lung injury and chronic repair leading to pulmonary fibrosis
J. Clin. Invest.
2001
, vol. 
107
 (pg. 
1529
-
1536
)
[PubMed]
51
Sime
 
P.J.
Marr
 
R.A.
Gauldie
 
D.
Xing
 
Z.
Hewlett
 
B.R.
Graham
 
F.L.
Gauldie
 
J.
 
Transfer of tumor necrosis factor-alpha to rat lung induces severe pulmonary inflammation and patchy interstitial fibrogenesis with induction of transforming growth factor-beta1 and myofibroblasts
Am. J. Pathol.
1998
, vol. 
153
 (pg. 
825
-
832
)
[PubMed]
52
Lee
 
C.G.
Cho
 
S.J.
Kang
 
M.J.
Chapoval
 
S.P.
Lee
 
P.J.
Noble
 
P.W.
Yehualaeshet
 
T.
Lu
 
B.
Flavell
 
R.A.
Milbrandt
 
J.
, et al 
Early growth response gene 1-mediated apoptosis is essential for transforming growth factor beta1-induced pulmonary fibrosis
J. Exp. Med.
2004
, vol. 
200
 (pg. 
377
-
389
)
[PubMed]
53
Hardie
 
W.D.
Korfhagen
 
T.R.
Sartor
 
M.A.
Prestridge
 
A.
Medvedovic
 
M.
 
Le Cras
 
T.D.
Ikegami
 
M.
Wesselkamper
 
S.C.
Davidson
 
C.
Dietsch
 
M.
, et al 
Genomic profile of matrix and vasculature remodeling in TGF-alpha induced pulmonary fibrosis
Am. J. Respir. Cell Mol. Biol.
2007
, vol. 
37
 (pg. 
309
-
321
)
[PubMed]
54
Korfhagen
 
T.R.
Swantz
 
R.J.
Wert
 
S.E.
McCarty
 
J.M.
Kerlakian
 
C.B.
Glasser
 
S.W.
Whitsett
 
J.A.
 
Respiratory epithelial cell expression of human transforming growth factor-alpha induces lung fibrosis in transgenic mice
J. Clin. Invest.
1994
, vol. 
93
 (pg. 
1691
-
1699
)
[PubMed]
55
Lee
 
C.G.
Homer
 
R.J.
Zhu
 
Z.
Lanone
 
S.
Wang
 
X.
Koteliansky
 
V.
Shipley
 
J.M.
Gotwals
 
P.
Noble
 
P.
Chen
 
Q.
, et al 
Interleukin-13 induces tissue fibrosis by selectively stimulating and activating transforming growth factor beta(1)
J. Exp. Med.
2001
, vol. 
194
 (pg. 
809
-
821
)
[PubMed]
56
Pierce
 
E.M.
Carpenter
 
K.
Jakubzick
 
C.
Kunkel
 
S.L.
Flaherty
 
K.R.
Martinez
 
F.J.
Hogaboam
 
C.M.
 
Therapeutic targeting of CC ligand 21 or CC chemokine receptor 7 abrogates pulmonary fibrosis induced by the adoptive transfer of human pulmonary fibroblasts to immunodeficient mice
Am. J. Pathol.
2007
, vol. 
170
 (pg. 
1152
-
1164
)
[PubMed]
57
Trujillo
 
G.
Meneghin
 
A.
Flaherty
 
K.R.
Sholl
 
L.M.
Myers
 
J.L.
Kazerooni
 
E.A.
Gross
 
B.H.
Oak
 
S.R.
Coelho
 
A.L.
Evanoff
 
H.
, et al 
TLR9 differentiates rapidly from slowly progressing forms of idiopathic pulmonary fibrosis
Sci. Transl. Med.
2010
, vol. 
2
 pg. 
57ra82
 
[PubMed]
58
Sisson
 
T.H.
Mendez
 
M.
Choi
 
K.
Subbotina
 
N.
Courey
 
A.
Cunningham
 
A.
Dave
 
A.
Engelhardt
 
J.F.
Liu
 
X.
White
 
E.S.
, et al 
Targeted injury of type II alveolar epithelial cells induces pulmonary fibrosis
Am. J. Respir. Crit. Care Med.
2010
, vol. 
181
 (pg. 
254
-
263
)
[PubMed]
59
Syrja
 
P.
Heikkila
 
H.P.
Lilja-Maula
 
L.
Krafft
 
E.
Clercx
 
C.
Day
 
M.J.
Ronty
 
M.
Myllarniemi
 
M.
Rajamaki
 
M.M.
 
The histopathology of idiopathic pulmonary fibrosis in West Highland white terriers shares features of both non-specific interstitial pneumonia and usual interstitial pneumonia in man
J. Comp. Pathol.
2013
, vol. 
149
 (pg. 
303
-
313
)
[PubMed]
60
Miele
 
A.
Dhaliwal
 
K.
Du
 
T.N.
Murchison
 
J.T.
Dhaliwal
 
C.
Brooks
 
H.
Smith
 
S.H.
Hirani
 
N.
Schwarz
 
T.
Haslett
 
C.
, et al 
Chronic pleuropulmonary fibrosis and elastosis of aged donkeys—similarities to human pleuroparenchymal fibroelastosis (PPFE)
Chest
2014
 
doi: 10.1378/chest.13-1306
61
Burger
 
R.M.
Peisach
 
J.
Horwitz
 
S.B.
 
Activated bleomycin. A transient complex of drug, iron, and oxygen that degrades DNA
J. Biol. Chem.
1981
, vol. 
256
 (pg. 
11636
-
11644
)
[PubMed]
62
Scotton
 
C.J.
Chambers
 
R.C.
 
Bleomycin revisited: towards a more representative model of IPF?
Am. J. Physiol. Lung Cell Mol. Physiol.
2010
, vol. 
299
 (pg. 
L439
-
L441
)
[PubMed]
63
Matute-Bello
 
G.
Frevert
 
C.W.
Martin
 
T.R.
 
Animal models of acute lung injury
Am. J. Physiol. Lung Cell Mol. Physiol.
2008
, vol. 
295
 (pg. 
L379
-
L399
)
[PubMed]
64
Degryse
 
A.L.
Tanjore
 
H.
Xu
 
X.C.
Polosukhin
 
V.V.
Jones
 
B.R.
McMahon
 
F.B.
Gleaves
 
L.A.
Blackwell
 
T.S.
Lawson
 
W.E.
 
Repetitive intratracheal bleomycin models several features of idiopathic pulmonary fibrosis
Am. J. Physiol. Lung Cell Mol. Physiol.
2010
, vol. 
299
 (pg. 
L442
-
L452
)
[PubMed]
65
Ask
 
K.
Labiris
 
R.
Farkas
 
L.
Moeller
 
A.
Froese
 
A.
Farncombe
 
T.
McClelland
 
G.B.
Inman
 
M.
Gauldie
 
J.
Kolb
 
M.R.
 
Comparison between conventional and “clinical” assessment of experimental lung fibrosis
J. Transl. Med.
2008
, vol. 
6
 pg. 
16
 
[PubMed]
66
Babin
 
A.L.
Cannet
 
C.
Gerard
 
C.
Saint-Mezard
 
P.
Page
 
C.P.
Sparrer
 
H.
Matsuguchi
 
T.
Beckmann
 
N.
 
Bleomycin-induced lung injury in mice investigated by MRI: model assessment for target analysis
Magn. Reson. Med.
2012
, vol. 
67
 (pg. 
499
-
509
)
[PubMed]
67
Choi
 
E.J.
Jin
 
G.Y.
Bok
 
S.M.
Han
 
Y.M.
Lee
 
Y.S.
Jung
 
M.J.
Kwon
 
K.S.
 
Serial micro-CT assessment of the therapeutic effects of rosiglitazone in a bleomycin-induced lung fibrosis mouse model
Korean J. Radiol.
2014
, vol. 
15
 (pg. 
448
-
455
)
[PubMed]
68
Lee
 
H.J.
Goo
 
J.M.
Kim
 
N.R.
Kim
 
M.A.
Chung
 
D.H.
Son
 
K.R.
Kim
 
H.C.
Lee
 
C.H.
Park
 
C.M.
Chun
 
E.J.
Im
 
J.G.
 
Semiquantitative measurement of murine bleomycin-induced lung fibrosis in in vivo and postmortem conditions using microcomputed tomography: correlation with pathologic scores–initial results
Invest. Radiol.
2008
, vol. 
43
 (pg. 
453
-
460
)
[PubMed]
69
Scotton
 
C.J.
Hayes
 
B.
Alexander
 
R.
Datta
 
A.
Forty
 
E.J.
Mercer
 
P.F.
Blanchard
 
A.
Chambers
 
R.C.
 
Ex vivo micro-computed tomography analysis of bleomycin-induced lung fibrosis for preclinical drug evaluation
Eur. Respir. J.
2013
, vol. 
42
 (pg. 
1633
-
1645
)
[PubMed]
70
Kakugawa
 
T.
Mukae
 
H.
Hayashi
 
T.
Ishii
 
H.
Abe
 
K.
Fujii
 
T.
Oku
 
H.
Miyazaki
 
M.
Kadota
 
J.
Kohno
 
S.
 
Pirfenidone attenuates expression of HSP47 in murine bleomycin-induced pulmonary fibrosis
Eur. Respir. J.
2004
, vol. 
24
 (pg. 
57
-
65
)
[PubMed]
71
Chaudhary
 
N.I.
Roth
 
G.J.
Hilberg
 
F.
Muller-Quernheim
 
J.
Prasse
 
A.
Zissel
 
G.
Schnapp
 
A.
Park
 
J.E.
 
Inhibition of PDGF, VEGF and FGF signalling attenuates fibrosis
Eur. Respir. J.
2007
, vol. 
29
 (pg. 
976
-
985
)
[PubMed]
72
Aono
 
Y.
Nishioka
 
Y.
Inayama
 
M.
Ugai
 
M.
Kishi
 
J.
Uehara
 
H.
Izumi
 
K.
Sone
 
S.
 
Imatinib as a novel antifibrotic agent in bleomycin-induced pulmonary fibrosis in mice
Am. J. Respir. Crit. Care Med.
2005
, vol. 
171
 (pg. 
1279
-
1285
)
[PubMed]
73
Chaudhary
 
N.I.
Schnapp
 
A.
Park
 
J.E.
 
Pharmacologic differentiation of inflammation and fibrosis in the rat bleomycin model
Am. J. Respir. Crit. Care Med.
2006
, vol. 
173
 (pg. 
769
-
776
)
[PubMed]
74
Ashley
 
S.L.
Jegal
 
Y.
Moore
 
T.A.
van Dyk
 
L.F.
Laouar
 
Y.
Moore
 
B.B.
 
gamma-Herpes virus-68, but not Pseudomonas aeruginosa or influenza A (H1N1), exacerbates established murine lung fibrosis
Am. J. Physiol. Lung Cell Mol. Physiol.
2014
, vol. 
307
 (pg. 
L219
-
L230
)
[PubMed]
75
Redente
 
E.F.
Jacobsen
 
K.M.
Solomon
 
J.J.
Lara
 
A.R.
Faubel
 
S.
Keith
 
R.C.
Henson
 
P.M.
Downey
 
G.P.
Riches
 
D.W.
 
Age and sex dimorphisms contribute to the severity of bleomycin-induced lung injury and fibrosis
Am. J. Physiol. Lung Cell Mol. Physiol.
2011
, vol. 
301
 (pg. 
L510
-
L518
)
[PubMed]
76
Sueblinvong
 
V.
Neujahr
 
D.C.
Mills
 
S.T.
Roser-Page
 
S.
Ritzenthaler
 
J.D.
Guidot
 
D.
Rojas
 
M.
Roman
 
J.
 
Predisposition for disrepair in the aged lung
Am. J. Med. Sci.
2012
, vol. 
344
 (pg. 
41
-
51
)
[PubMed]
77
Burge
 
S.
Wedzicha
 
J.A.
 
COPD exacerbations: definitions and classifications
Eur. Respir. J. Suppl.
2003
, vol. 
41
 (pg. 
46s
-
53s
)
[PubMed]
78
Papi
 
A.
Bellettato
 
C.M.
Braccioni
 
F.
Romagnoli
 
M.
Casolari
 
P.
Caramori
 
G.
Fabbri
 
L.M.
Johnston
 
S.L.
 
Infections and airway inflammation in chronic obstructive pulmonary disease severe exacerbations
Am. J. Respir. Crit. Care Med.
2006
, vol. 
173
 (pg. 
1114
-
1121
)
[PubMed]
79
Sethi
 
S.
Murphy
 
T.F.
 
Bacterial infection in chronic obstructive pulmonary disease in 2000: a state-of-the-art review
Clin. Microbiol. Rev.
2001
, vol. 
14
 (pg. 
336
-
363
)
[PubMed]
80
Falsey
 
A.R.
Formica
 
M.A.
Hennessey
 
P.A.
Criddle
 
M.M.
Sullender
 
W.M.
Walsh
 
E.E.
 
Detection of respiratory syncytial virus in adults with chronic obstructive pulmonary disease
Am. J. Respir. Crit. Care Med.
2006
, vol. 
173
 (pg. 
639
-
643
)
[PubMed]
81
Hamelin
 
M.E.
Boivin
 
G.
 
Human metapneumovirus: a ubiquitous and long-standing respiratory pathogen
Pediatr. Infect. Dis. J.
2005
, vol. 
24
 (pg. 
S203
-
S207
)
[PubMed]
82
Rohde
 
G.
Wiethege
 
A.
Borg
 
I.
Kauth
 
M.
Bauer
 
T.T.
Gillissen
 
A.
Bufe
 
A.
Schultze-Werninghaus
 
G.
 
Respiratory viruses in exacerbations of chronic obstructive pulmonary disease requiring hospitalisation: a case-control study
Thorax
2003
, vol. 
58
 (pg. 
37
-
42
)
[PubMed]
83
Aaron
 
S.D.
Angel
 
J.B.
Lunau
 
M.
Wright
 
K.
Fex
 
C.
Le
 
S.N.
Dales
 
R.E.
 
Granulocyte inflammatory markers and airway infection during acute exacerbation of chronic obstructive pulmonary disease
Am. J. Respir. Crit. Care Med.
2001
, vol. 
163
 (pg. 
349
-
355
)
[PubMed]
84
Jackson
 
D.J.
Johnston
 
S.L.
 
The role of viruses in acute exacerbations of asthma
J. Allergy Clin. Immunol.
2010
, vol. 
125
 (pg. 
1178
-
1187
)
[PubMed]
85
Clarke
 
D.L.
Davis
 
N.H.
Majithiya
 
J.B.
Piper
 
S.C.
Lewis
 
A.
Sleeman
 
M.A.
Corkill
 
D.J.
May
 
R.D.
 
Development of a mouse model mimicking key aspects of a viral asthma exacerbation
Clin. Sci.
2014
, vol. 
126
 (pg. 
567
-
580
)
[PubMed]
86
Ford
 
W.R.
Blair
 
A.E.
Evans
 
R.L.
John
 
E.
Bugert
 
J.J.
Broadley
 
K.J.
Kidd
 
E.J.
 
Human parainfluenza type 3 virus impairs the efficacy of glucocorticoids to limit allergy-induced pulmonary inflammation in guinea-pigs
Clin. Sci.
2013
, vol. 
125
 (pg. 
471
-
482
)
[PubMed]
87
Mori
 
H.
Parker
 
N.S.
Rodrigues
 
D.
Hulland
 
K.
Chappell
 
D.
Hincks
 
J.S.
Bright
 
H.
Evans
 
S.M.
Lamb
 
D.J.
 
Differences in respiratory syncytial virus and influenza infection in a house-dust-mite-induced asthma mouse model: consequences for steroid sensitivity
Clin. Sci.
2013
, vol. 
125
 (pg. 
565
-
574
)
[PubMed]
88
Rochlitzer
 
S.
Hoymann
 
H.G.
Muller
 
M.
Braun
 
A.
 
No exacerbation but impaired anti-viral mechanisms in a rhinovirus-chronic allergic asthma mouse model
Clin. Sci.
2014
, vol. 
126
 (pg. 
55
-
65
)
[PubMed]
89
Gualano
 
R.C.
Hansen
 
M.J.
Vlahos
 
R.
Jones
 
J.E.
Park-Jones
 
R.A.
Deliyannis
 
G.
Turner
 
S.J.
Duca
 
K.A.
Anderson
 
G.P.
 
Cigarette smoke worsens lung inflammation and impairs resolution of influenza infection in mice
Respir. Res.
2008
, vol. 
9
 pg. 
53
 
[PubMed]
90
Bauer
 
C.M.
Zavitz
 
C.C.
Botelho
 
F.M.
Lambert
 
K.N.
Brown
 
E.G.
Mossman
 
K.L.
Taylor
 
J.D.
Stampfli
 
M.R.
 
Treating viral exacerbations of chronic obstructive pulmonary disease: insights from a mouse model of cigarette smoke and H1N1 influenza infection
PLoS One
2010
, vol. 
5
 pg. 
e13251
 
[PubMed]
91
Robbins
 
C.S.
Bauer
 
C.M.
Vujicic
 
N.
Gaschler
 
G.J.
Lichty
 
B.D.
Brown
 
E.G.
Stampfli
 
M.R.
 
Cigarette smoke impacts immune inflammatory responses to influenza in mice
Am. J. Respir. Crit. Care Med.
2006
, vol. 
174
 (pg. 
1342
-
1351
)
[PubMed]
92
Krimmer
 
D.I.
Oliver
 
B.G.
 
What can in vitro models of COPD tell us?
Pulm. Pharmacol. Ther.
2011
, vol. 
24
 (pg. 
471
-
477
)
[PubMed]
93
McAnulty
 
R.J.
 
Models and approaches to understand the role of airway remodelling in disease
Pulm. Pharmacol. Ther.
2011
, vol. 
24
 (pg. 
478
-
486
)
[PubMed]
94
Meurs
 
H.
Gosens
 
R.
Zaagsma
 
J.
 
Airway hyperresponsiveness in asthma: lessons from in vitro model systems and animal models
Eur. Respir. J.
2008
, vol. 
32
 (pg. 
487
-
502
)
[PubMed]
95
Booth
 
A.J.
Hadley
 
R.
Cornett
 
A.M.
Dreffs
 
A.A.
Matthes
 
S.A.
Tsui
 
J.L.
Weiss
 
K.
Horowitz
 
J.C.
Fiore
 
V.F.
Barker
 
T.H.
, et al 
Acellular normal and fibrotic human lung matrices as a culture system for in vitro investigation
Am. J. Respir. Crit. Care Med.
2012
, vol. 
186
 (pg. 
866
-
876
)
[PubMed]
96
Morin
 
J.P.
Baste
 
J.M.
Gay
 
A.
Crochemore
 
C.
Corbiere
 
C.
Monteil
 
C.
 
Precision cut lung slices as an efficient tool for in vitro lung physio-pharmacotoxicology studies
Xenobiotica
2013
, vol. 
43
 (pg. 
63
-
72
)
[PubMed]
97
Sanderson
 
M.J.
 
Exploring lung physiology in health and disease with lung slices
Pulm. Pharmacol. Ther.
2011
, vol. 
24
 (pg. 
452
-
465
)
[PubMed]
98
Wheelock
 
C.E.
Goss
 
V.M.
Balgoma
 
D.
Nicholas
 
B.
Brandsma
 
J.
Skipp
 
P.J.
Snowden
 
S.
Burg
 
D.
D’Amico
 
A.
Horvath
 
I.
, et al 
Application of ‘omics technologies to biomarker discovery in inflammatory lung diseases
Eur. Respir. J.
2013
, vol. 
42
 (pg. 
802
-
825
)
[PubMed]
99
Westra
 
I.M.
Pham
 
B.T.
Groothuis
 
G.M.
Olinga
 
P.
 
Evaluation of fibrosis in precision-cut tissue slices
Xenobiotica
2013
, vol. 
43
 (pg. 
98
-
112
)
[PubMed]
100
Slack
 
R.J.
Barrett
 
V.J.
Morrison
 
V.S.
Sturton
 
R.G.
Emmons
 
A.J.
Ford
 
A.J.
Knowles
 
R.G.
 
In vitro pharmacological characterization of vilanterol, a novel long-acting beta2-adrenoceptor agonist with 24-hour duration of action
J. Pharmacol. Exp. Ther.
2013
, vol. 
344
 (pg. 
218
-
230
)
[PubMed]
101
Ulrich
 
K.
Hincks
 
J.S.
Walsh
 
R.
Wetterstrand
 
E.M.
Fidock
 
M.D.
Sreckovic
 
S.
Lamb
 
D.J.
Douglas
 
G.J.
Yeadon
 
M.
Perros-Huguet
 
C.
Evans
 
S.M.
 
Anti-inflammatory modulation of chronic airway inflammation in the murine house dust mite model
Pulm. Pharmacol. Ther.
2008
, vol. 
21
 (pg. 
637
-
647
)
[PubMed]
102
Wakahara
 
K.
Tanaka
 
H.
Takahashi
 
G.
Tamari
 
M.
Nasu
 
R.
Toyohara
 
T.
Takano
 
H.
Saito
 
S.
Inagaki
 
N.
Shimokata
 
K.
Nagai
 
H.
 
Repeated instillations of Dermatophagoides farinae into the airways can induce Th2-dependent airway hyperresponsiveness, eosinophilia and remodeling in mice: effect of intratracheal treatment of fluticasone propionate
Eur. J. Pharmacol.
2008
, vol. 
578
 (pg. 
87
-
96
)
[PubMed]
103
Muhlhausler
 
B.S.
Bloomfield
 
F.H.
Gillman
 
M.W.
 
Whole animal experiments should be more like human randomized controlled trials
PLoS Biol.
2013
, vol. 
11
 pg. 
e1001481
 
[PubMed]
104
Kane
 
R.L.
Wang
 
J.
Garrard
 
J.
 
Reporting in randomized clinical trials improved after adoption of the CONSORT statement
J. Clin. Epidemiol.
2007
, vol. 
60
 (pg. 
241
-
249
)
[PubMed]
105
Kilkenny
 
C.
Browne
 
W.J.
Cuthill
 
I.C.
Emerson
 
M.
Altman
 
D.G.
 
Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research
Osteoarthritis Cartilage
2012
, vol. 
20
 (pg. 
256
-
260
)
[PubMed]
106
Henderson
 
V.C.
Kimmelman
 
J.
Fergusson
 
D.
Grimshaw
 
J.M.
Hackam
 
D.G.
 
Threats to validity in the design and conduct of preclinical efficacy studies: a systematic review of guidelines for in vivo animal experiments
PLoS Med.
2013
, vol. 
10
 pg. 
e1001489
 
[PubMed]
107
Couzin-Frankel
 
J.
 
When mice mislead
Science
2013
, vol. 
342
 (pg. 
922
-
923
925
[PubMed]