© 2006 American Thoracic Society doi: 10.1164/rccm.2609002
The Mouse TrapIt Still Yields Few Answers in AsthmaUniversity of Pittsburgh Medical Center, Pittsburgh, Pennsylvania
University of Southampton, Southampton, United Kingdom The discovery and "re-discovery" that asthma is an inflammatory disease, arguably the most important advance in the understanding of this disease, came about through human investigations (1, 2). Although details have improved, large gaps remain in our understanding of the onset, persistence, consequences, and treatment of this inflammation. Since 1966, there have been over 8,000 papers published using human subjects, tissue, or cells to study asthma immunology and pathophysiology. In that period, over 2,300 papers were published using murine models to study allergic airway pathobiology. The majority (60%) of murine studies have been published within the last 5 years, indicating an enormous escalation in their use. In fact, most prominent basic laboratories use some variation of this murine system. In contrast, only 22% of human-based asthma studies have been published within the last 5 years. The best murine studies are published in the highest impact journals, such as Nature and Science, yet only the most extraordinary human-based studies receive such recognition. Although murine models have provided insights into immunologic processes, have they really improved the understanding of human asthma, or, more importantly for patients, led to the discovery of new therapies? Generally speaking, the answer has to be "no." Furthermore, there are several instances where animal "models" of asthma may have even misdirected resources and thinking. Despite their attractiveness from a technologic, transgenic, and drug-screening perspective, murine models have a range of shortcomings, many of which were outlined several years ago in Trends in Pharmacological Sciences (3). Those "original" defects remain unresolved and more are now appreciated. Until these shortcomings are addressed, murine models will continue to limit the understanding and treatment of human asthma. CHRONIC ASTHMA IS A DISEASE UNIQUE TO HUMANS Mice do not have asthma. Even the most "hyperresponsive" strains of mice do not exhibit spontaneous "symptoms" consistent with asthma, nor is their bronchial hyperresponsiveness (BHR) "severe" or understood mechanistically, as most mouse airways do not contain smooth muscle bundles. Although mice can be manipulated to develop allergic/Th2-type immune responses, this usually requires highly artificial peritoneal allergen sensitization procedures and adjunctive (alum) stimulation. The murine allergen challenge system does not involve either immediate or late airway obstruction, which can be physiologically measured without an additional (methacholine) challenge. Although immunization methods have been developed to invoke long-term allergen sensitivity, as opposed to tolerance, we are not aware of any murine model that develops spontaneous acute bronchoconstriction which characterizes human asthma (4). Almost total dependency on allergen-driven murine models also diminishes the importance of other environmental factors to the development of asthma, such as oxidant stress, viral infection, obesity and other aspects of diet, and exposure to tobacco smoke and occupational and other pollutants. In human asthma, the population-attributable risk (PAR) of atopy on asthma is less than 50% (5), and for Der p1 and Fel d1 exposure, the PAR is only 4 and 11%, respectively (6), raising questions whether atopy/allergy has been overinterpreted as a risk factor for the origins of asthma. Studies describing IgE and Th2-type lymphocytic responses in mice should not refer to them as models of "asthma," where the disease focus is the conducting airway, but as models of systemic allergic sensitization, which are frequently self-limiting (7). ASTHMA COMMONLY ORIGINATES IN EARLY LIFE It is likely that differences between mouse and human disease begin early in the development of the species. Mouse lungs are fully developed at birth, whereas both nonhuman and human primate lungs continue to develop from months to years after birth. In asthma, which often begins in the first few months of life, this difference has considerable implications. In humans, a very different, immature lung is exposed to the environment during times of initial exposures to antigens or agents that interfere with lung maturation (e.g., pollutants and diet), as compared with the mature lung present in the mouse. This immature human lung, both at a structural and immunologic level, responds much differently from that of a mouse lung. Examples of this are the effect of maternal smoking in pregnancy to impair lung growth (8) and increase asthma and BHR in the offspring (9). CHRONIC ASTHMA IS A DISORDER THAT INVOLVES CHANGES IN THE EPITHELIAL–MESENCHYMAL UNIT SPECIFIC TO HUMAN LUNGS Even in adulthood, the structure of mouse airways bears little resemblance to that in humans (Figure 1). There are many fewer airway generations in mice (3, 4), as compared with the 27 to 29 generations in humans, suggesting much greater variety in airway structure in the latter, as well as differences in flow dynamics. Mouse lungs lack submucosal glands in every airway except the trachea, whereas human airways have an abundance of these glands throughout the large and medium-sized airways. Furthermore, in the large airways of the mouse, the bronchial epithelium is not fully stratified as it is in the larger airways in humans. This may have an impact on the way the airway epithelium responds to different inhaled stimuli (10), such as respiratory viruses and pollutants (11).
In humans, the smooth muscle is arranged in spiral bundles, which in asthma increase markedly in size due to muscle cell hypertrophy, hyperplasia, or both (12). The "smooth muscle" in mouse airways beyond the trachea and first-generation bronchi has the appearance of myofibroblasts rather than true smooth muscle cells (13). The role of these cells in the reported BHR in mice is being questioned (14). HUMAN ASTHMA IS RESTRICTED TO THE AIRWAYS The majority of "allergic" inflammation in mouse airways is parenchymal and vascular rather than generally restricted to conducting airways, as it is in human asthma (4). The primary nasal breathing pattern, airway structure and flow dynamics, sensitization, and challenge procedures in the mouse predispose to this nasal and alveolar response rather than one directed to the lower conducting airways (15). Whether the vascular inflammation comes from systemic absorption, the lack of a bronchial circulation in murine lungs, or peculiar local trafficking of antigens is not known (16). The antigen provocation model in mice more resembles a systemic immune complex response with pulmonary vasculitis. This unusual pathology is rarely discussed in papers describing murine "asthma models." In humans, the pathogenesis of eosinophilic diseases with vascular inflammation (e.g., Churg–Strauss syndrome, polyarteritis nodosa) is quite different from that of the classic localized airway inflammation seen in most cases of asthma. AIRWAY PHYSIOLOGY IS DIFFICULT TO MEASURE IN MICE Airway obstruction is difficult to measure in living/nonsedated mice. PenH (enhanced pause), although widely used as an index of airway obstruction, likely is more a measure of distress of the animal than actual obstruction (17). Clearly, allergic responses could cause "distress" in a variety of ways. Unrestrained plethysmography is also problematic (18, 19). The alternative, sedating and tracheostomizing the animal to perform airway resistance measures, brings in additional factors related to anesthesia that may or may not accurately reflect airway resistance in nonsedated animals. MICE EXHIBIT A POLARIZED T-LYMPHOCYTE RESPONSE
Th1/Th2 immune pathways in mice are highly distinct. In contrast, it is almost impossible to describe a pure Th1 or Th2 response in humans. IFN- TARGET CELLS MAY DIFFER IN RESPONSES Tissue/cellular responses related to particular cytokines also differ between mice and humans. IL-13 elicits stronger relative responses in murine structural cells than those observed in human airway cells. Small amounts of IL-13 are highly fibrogenic in mice, but much less so in humans (27). In addition, synergies between IL-13 and transforming growth factor (TGF)-ß to induce large amounts of eotaxin from human fibroblasts are absent, and even in opposition in murine fibroblasts (28). The addition of TGF-ß blunts an already high level of eotaxin production to IL-13 in murine fibroblasts (S. Wenzel, unpublished data). CHRONIC ALLERGEN MODELS IN MICE DO NOT REPLICATE STRUCTURAL CHANGES ASSOCIATED WITH REMODELING IN HUMAN ASTHMA The "newest" addition to mouse models is the development of "chronic" protocols purported to better represent human asthma (29, 30). The features of these models reportedly include long-term effects on hyperresponsiveness and structural changes. In some cases, lung parenchymal and vascular inflammation/remodeling are also noted, which are not seen in human asthma (16, 31, 32). However, one of the most characteristic features of chronic human asthma, large increases in airway smooth muscle, does not occur (33). The term "remodeling" in the mouse studies often describes the subepithelial deposition of matrix proteins. This lamina reticularis, the increased matrix deposition in the subepithelial basement membrane, is of questionable importance in human asthma because it occurs in allergic rhinitis in the absence of asthma and in eosinophilic bronchitis in the absence of BHR (34). In human asthma, airway remodeling extends beyond this to include the full thickness of the airway wall and the region of alveolar attachments (33, 35, 36). Furthermore, in most cases, remodeling elements of murine models are responsive to corticosteroids, yet the response in humans has been much less (37–40). We are not aware of murine models that demonstrate the chronic increase in lung volumes and compliance that is representative of more severe human asthma (35, 41, 42). HUMANS ARE "OUTBRED" Research mice are "inbred" strains. For whatever hypothesis being addressed, a mouse strain is often found in support of the hypothesis. Mice vary in their underlying airway responsiveness and in their ability to develop "allergic" immunologic responses. BALB/c and A/J mice are known to respond to immunization and sensitization protocols with a profound allergic/inflammatory response (43). In contrast, C57Bl/6 mice respond to the same protocol with a much smaller response (43). When mice are chosen for transgenic/knockout studies, the strain chosen is likely to impact the results. Indeed, genetic background is being recognized as crucial to the phenotype experienced in single-gene–manipulated mice. A clear example of this is the "twin" studies (using two different mouse strains) recently reported in Science, in which eosinophils were selectively removed. Although both were reported to prove the importance (or not) of the eosinophil, the results reported nearly opposite findings (44, 45). Applying classic genetic studies to these mouse strains would seem to be imperative. PROVING THE IMPORTANCE OF A PATHWAY IN A MOUSE DOES NOT PROVE THE IMPORTANCE OF THE SAME PATHWAY IN HUMAN ASTHMA Most transgenic mouse studies are only able to prove the importance of a particular mediator or pathway in a particular mouse, using a specific immunization/sensitization protocol. The transgenic and knockout technologies themselves may have broader impact than the pathway targeted, depending on when the pathway intervention is initiated. Large numbers of studies have overexpressed or knocked out a particular molecule. In many cases, these single interventions alter allergic responses. These interventions are then picked up by pharmaceutical companies and more animal studies are done. Unfortunately, in the majority of cases (e.g., very late antigen-4 [VLA-4], IL-4, IL-5, bradykinin, neurokinins, platelet activating factor), they fail to work in human asthma (23, 46). For example, the many studies of IL-5 in mice would have predicted a high degree of efficacy in humans, yet this was not the case (24). Only three targets (leukotrienes [LTs], IgE, and phosphodiesterase 4) involved in airway responses in mice have translated into clinical benefits for humans. With each of these pathways, there was abundant human-based evidence that these were likely to be important. In contrast, in the case of anti-IgE (omalizumab), the efficacy of the response in mice is highly dependent on the method used to sensitize the animal (47, 48). CONCLUSIONS Certainly, use of human subjects for the study of human asthma is not without problems. Only a limited number of approaches are ethically acceptable. Humans almost always have the disease before they are studied, such that dynamic studies of asthma development are difficult (although with occupational asthma, not impossible). Human variability, due to that mixed genetic background described above, often makes results difficult to interpret but represents the very basis of why a disease occurs in one person and not another. However, there are advantages to human research as well. Confirmed results in humans are almost certain to have some application to human disease. If one replicates most or all of Koch's postulates in human subjects for a given mediator (e.g., cysteinyl LTs), the findings will likely be of direct relevance to human disease. Although we are not advocating an abolition of mouse studies/models, taking an observation in humans "back" to a mouse model for "confirmatory" studies will likely distract from the primary goal, improvement of human asthma. Once a target (identified in human or mouse studies) has shown potential in mice, it should be quickly studied in humans, first on an in vitro level and then, as warranted, in experimental medicine and therapeutic trials. Translational medicine should and must operate in both directions. Human disease informs more basic research including animal models and vice versa. Basic pathologic and immunologic studies in humans, with all of their many "issues," deserve to return to the forefront of efforts to understand this complex human disease. When mouse models are used, publications describing their results must include reference to "mouse" in the title and not simply "asthma." Furthermore, the limitations of these systems must be discussed with respect to human asthma. As emphasized recently by numerous scientists, every effort should be made to improve these murine systems such that they more closely reflect the disease for which they claim to be models (7, 49). Acknowledgments The authors thank Ms. Roz Dudden and Ms. Shandra Protzko for their assistance in researching the numbers of mouse and human trials in asthma or allergic models. They also thank Ms. Shirley Pearce for her technical assistance and Drs. Hong Wei Chu and Silvana Balzar for the pathology figures. FOOTNOTES Conflict of Interest Statement: Neither author has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
and interleukin (IL)-12 are found in human asthma, sometimes in high amounts (
(a Th1/innate-type cytokine) looks more promising (
