| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Vermont Lung Center, Department of Pulmonary and Critical Care Medicine, University of Vermont College of Medicine, Burlington, Vermont
Submitted 2 November 2005 ; accepted in final form 21 September 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
computational model; airway closure; asthma; mouse models of asthma
Mice are often used as in vivo models of asthma as they allow for the manipulation of genetics and for the detailed study of molecular and structural changes induced by inflammation. Because mice can be sedated and tracheostomized, eliminating the effects of voluntary skeletal muscle movement, as well as upper airway obstruction and shunting, their use allows one to conduct sensitive physiological studies (1). Using a forced oscillation technique and a computational model of the lung, we have recently shown that in a commonly used mouse model of asthma, BALB/c mice with inflammation of the airways induced by antigen sensitization and challenge, AHR can be attributed to airway wall thickening and exaggerated airway closure (45). This finding is contrary to the commonly held notion that AHR is attributable to alterations in airway smooth muscle function.
Multiple studies have been published that make the assumption that AHR is physiologically the same in different mouse models of asthma (6, 7, 11, 12, 26, 27). In the present study, we sought to apply an approach similar to the one we used previously (45) to another commonly used mouse model of asthma that is known to have intrinsic AHR, A/J mice (26). This intrinsic AHR in A/J mice has been attributed to the presence of airway smooth muscle with an increased velocity of shortening (9). We chose to study A/J mice in the absence of airway inflammation and to compare the physiology of their intrinsic AHR to that of mice that had undergone antigen sensitization and challenge. We hypothesized that AHR in the A/J mice would be physiologically distinct from the AHR in antigen-sensitized and -challenged mice.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Antigen sensitization and challenge. BALB/c mice were injected intraperitoneally on days 0 and 14 with a 100-µl mixture of 50 µl ovalbumin (200 µl/ml) and 50 µl of alum (adjuvant). Beginning on day 21, the mice received daily aerosol challenges with ovalbumin for each of 3 days and were studied 48 h after the last challenge. Bronchoalveolar lavage cell counts and differentials were determined to confirm the presence or absence of inflammation.
AHR.
Mice were anesthetized with an intraperitoneal injection of pentobarbital sodium (90 mg/kg), tracheostomized, and then connected to a mechanical ventilator (flexiVent, Scireq, Montreal, Canada) via an 18-gauge intratracheal cannula. The mice were ventilated at 200 breaths/min with a delivered tidal volume of 0.20 ml against a positive end-expiratory pressure (PEEP) of 3 cmH2O applied by a water trap. Aerosolized methacholine challenges (untreated BALB/c n = 8, antigen-sensitized and -challenged BALB/c n = 10, A/J mice n = 9) were performed by delivering three successive concentrations of methacholine: 3.125, 12.5, and 50 mg/ml. Intravenous methacholine challenges (untreated BALB/c n = 6, antigen-sensitized and -challenged BALB/c n = 7, A/J mice n = 6) were performed by intravenously delivering, through a Silastic catheter that was surgically inserted into the left external jugular vein, six successive doses of methacholine: 0.0027, 0.0082, 0.0247, 0.0741, 0.222, and 0.666 mg/ml, where the volume injected was 1.85 µl/g body wt of mouse (20). Following each aerosol or intravenous methacholine challenge, every 10-s ventilation was interrupted allowing for a 1-s passive expiration followed by a 2-s broad-band (1–19.625 Hz) volume perturbation. The peak-to-peak excursion of the ventilator piston during delivery of these perturbations was 0.17 ml, resulting in a volume delivered of 
0.14 ml after accounting for gas compression in the ventilator cylinder and connecting tubing. The pressure and flow data obtained during application of the volume perturbations were used to calculate a complex input impedance (Zrs) of the respiratory system. Using an iterative scheme (45), Zrs was then fit to a uniformly ventilated model of the lung with a constant-phase tissue impedance described by (15):
 | (1) |
where RN is a Newtonian resistance composed mostly of the flow resistance of the conducting pulmonary airways, Iaw is the inertance of the gas in the central airways, G (tissue damping) reflects viscous dissipation of energy in the respiratory tissues as well as airflow heterogeneity, H (elastance) reflects elastic energy storage in the tissues, f is frequency, i = 
, and 
couples G and H. Using the frequency normalization scheme of Ito et al. (21), RN, G, and H all have units of centimeters H2O times seconds per milliliter. Linear interpolation of the methacholine dose-response curve was used to calculate the concentration of methacholine that would cause a 200% increase in each RN and G (PC200) or a 50% increase in H (PC50) for mice that received aerosolized methacholine. A 300% increase in RN and G (PC300) and a 30% increase in H (PC30) were used for mice that received intravenous methacholine. The PC values were chosen according to the appearance of the dose-response curves, so as to provide a point of analysis that represented both the sensitivity and maximal response of the curves in question. If for a given mouse the PC value was not reached, the maximal dose given was used as the PC value.
Computational model experiments.
The time course of bronchoconstriction was plotted for the 12.5 mg/ml aerosol and the 0.0247 mg/ml intravenous dose of methacholine. As previously described (45), we simulated the time course of bronchoconstriction using a computational model of the mouse lung that followed the asymmetric branching scheme of Horsfield et al. (17) with structural data for the mouse airway tree as described by Gomes and Bates (14). The airway structural data include the diameters and lengths of 19 airway orders, with a Horsfield delta value of 6 for orders 1–12 and delta values decreasing stepwise to zero for subsequent orders. Random values were assigned to the baseline (unconstricted) radii of the airways in the model, according to Gaussian distributions, having means and SDs appropriate to the airway order as per the anatomic data (14). The impedance of each airway (Zaw) in the model was calculated, assuming Poiseuille flow to determine resistance and the mass of the airway gas to determine inertance, as follows
 | (2) |
where r is airway radius, L is airway length, µ is gas viscosity, and 
is gas density. We neglected the influence of airway wall shunting by making the airway walls rigid. Each of the most distal airways terminated in an identical viscoelastic tissue unit with impedance given by the equation
 | (3) |
where Zti is tissue impedance, Gti is tissue damping G, and Hti is tissue H, with the ratio of Gti to Hti (Gti/Hti) being assigned a value of 0.1. The total impedance of the model (Zmod) was calculated by adding the individual Zaw and Zti in series or parallel, as appropriate. Zmod was calculated at each of the frequencies in the volume perturbation signal used to obtain Zrs experimentally in the mice. The Monte-Carlo approach was used to obtain a set of different Zmod by running the computational model multiple times, each time using a different statistical realization of the airway radii values. Simulations of the time course of bronchoconstriction were achieved by having the model airways transiently decrease their radii according to a prescribed function of time.
The time-course of fractional change in airway radius necessary to have the model simulate transient bronchoconstriction was determined as follows. First, the mean profile of the measured 
RN was added to unity, inverted, and the fourth root taken, as if to assume Poiseuille flow in the airways. Then, to force a good match between the simulated and measured RN profiles, the time course of fractional change in airway radius had to be further scaled by the empirically determined factor of 1.3 presumably because of the inherent heterogeneity in the model and the structural differences between a cast of an inflated fixed specimen and the lung of a living animal.
Visualization of airway closure. A mixed gas that is trapped behind a closed airway eventually will have all of the oxygen present extracted. With ambient air as the respired gas, this has little effect as oxygen comprises only 21% of ambient air. In contrast, if the fraction of inspired oxygen is increased to 100%, almost all the gas trapped behind closed airways will be extracted, over time resulting in collapse of all the alveoli distal to the closed airways with a consequent reduction in lung volume. This is a well-recognized phenomenon known as absorption atelectasis (13). In alveoli served by airways that remain open, the oxygen will be constantly replaced, and little absorption atelectasis will occur. Separate groups of mice that had undergone antigen sensitization and challenge were studied during ventilation with either humidified room air (n = 4) or humidified 100% oxygen (n = 4). For comparison, a group (n = 4) of A/J mice was studied during ventilation with humidified 100% oxygen. Once placed on the ventilator, as above, the anterior portion of the chest wall was removed. A single 12.5 mg/ml dose of aerosolized methacholine was delivered as above. At baseline and following methacholine, a series of high-resolution digital images of the lungs was captured each minute using a 1-megapixel black-and-white charge-coupled device camera. These images were taken after 1 s of passive expiration against 3 cmH2O PEEP and immediately before the start of the volume perturbations that were used to determine lung mechanics. The area of the visible lung was determined at baseline and 8 min following methacholine by tracing around the lung border using image analysis software (Image J, NIH, Bethesda, MD) (Fig. 1). The area of visible lung for each mouse and at each time point was measured independently by two observers blinded to the treatment and strain of mouse.
|
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Baseline values for all three parameters were similar for all six groups of mice (Table 1). Compared with controls, A/J mice and BALB/c mice that had undergone antigen sensitization and challenge were indeed hyperresponsive whether challenged with aerosol or intravenous methacholine (Figs. 2 and 3). However, while the antigen-sensitized and -challenged mice demonstrated AHR in all three parameters, i.e., RN, G, and H, A/J mice exhibited AHR that was similar in magnitude but limited to the parameters RN and G (Figs. 2 and 3).
|
|
|
|
RN, the simulated RN was virtually identical to the measured. The G simulated time course profile is also close to that of the measured G time course profile. However, the simulated and computational time course profiles for H are highly discordant (Fig. 5A). Thus, in the next simulation, we reduced the degree of airway closure by changing the closure threshold from 38 µm, as was used in the previous study (45), to 28 µm. With this modification we obtained an excellent match between computational and experimental time courses of RN, G, and H (Fig. 5B).
|
|
H is highly discordant with that of the experimental time course (Fig. 7A). When the model was further modified by constricting only the airways with a radius of 150 µm or greater and then further rescaling the fourth root of the experimental RN by a factor of 1.5 instead of 1.3, we achieved an improved fit (Fig. 7B).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Our study has certain limitations. The conclusions made herein are based on mathematical models of the lung, which embody assumptions that are always open to discussion. In the case of the model used to fit the impedance data (Eq. 1), the assumption of a particularly simple lung structure is necessary in order for the variables of the model to be uniquely identifiable from the data (15). Furthermore, the computational model (Eq. 2) is based on the airway tree structure of a strain of mouse (14) that is different from the strains used in the present study. Therefore, the accuracy of our computational model depends on the assumption that such scaling is valid. We are also assuming that it is appropriate to narrow all airways in the computational model by the same fraction to simulate bronchoconstriction, whereas in reality airways narrow heterogeneously (2, 28, 42), and heterogeneous narrowing has been used in a computational model of the lung (41). However, heterogeneity is incorporated in our model as the starting airway radii are randomly chosen. Nonetheless, because of the complex and interdependent nature of lung structure and function, isolating single components of the physiological response to methacholine is difficult to do without significantly altering their behavior, making studying components of bronchoconstriction in isolation nearly impossible (1). Therefore, we are left with having to infer the nature of the complete system on the basis of the extent to which we can reproduce its behavior using an anatomically based computational model. It has been recently pointed out that the inferences made with this approach are not unlike the inferences made in genomics (40).
Another limitation is that we assumed that measuring the visible surface area of the lung was a reasonable surrogate for total lung volume. It may be that areas that were not visible on the surface collapsed more. Nonetheless, at the very least we can conclude that the visible areas of the lung did or did not experience a reduction in volume. The removal of the chest wall could potentially reduce the outward recoil forces on the lung, thereby facilitating collapse of airways and alveoli. However, as it is known that the volume dependence of lung mechanics in mice is chest wall independent (39), a reasonable assumption is that the outward recoil of the chest wall is negligible in mice. Additionally, all mice in this arm of the study had the chest wall removed, making the comparison between groups valid. The strength of this approach is that it allows for direct visual confirmation of what was being inferred from the measurement of impedance and the computational modeling. In contrast to histological assessment of atelectasis, this technique allows for in vivo assessment that is not limited by fixation artifacts.
AHR in A/J mice has been studied previously. Quantitative trait locus mapping studies (6, 11, 12, 26) have linked genes to AHR when comparisons were made between A/J mice and a hyporesponsive strain of mice, C3HeJ. However, when antigen-sensitized and -challenged mice were used, different genetic loci were identified (11), which is perhaps not surprising in light of the results of the present study. Duguet et al. (9) also compared A/J mice to mouse strains with minimal and no AHR and found no apparent morphological difference between the strains. There was, however, a difference in the functional properties of the airway smooth muscle in that it had an increased velocity of shortening.
The physiological difference between A/J and antigen-sensitized and -challenged mice is likely the consequence of airway instability and closure. Airway closure will occur when small airways narrow to the point of complete apposition of the airway walls (41) or narrow enough that liquid bridging occurs across the airway lumen (29). If the exaggerated H response was purely a consequence of airway narrowing, we would expect to see it also in the A/J mice as the rise in RN, when aerosolized methacholine is used, is similar to, if not greater than, the rise in RN in antigen-sensitized and -challenged mice. This of course could also be the result of regional variation in aerosol deposition; however, when intravenous methacholine was administered the difference persisted, albeit at doses of methacholine that resulted in a greater response in RN compared with that which was seen with aerosolized methacholine. As there is no systemic circulation in the peripheral bronchioles of mice (32), a fact that is consistent with our computational model results, the exaggerated rise in H when intravenous methacholine is administered is either the result of large airways constricting to the point that they close, or an instability in the peripheral airways such that the diminished airflow secondary to large airway narrowing leads to their collapse. Even if the former is true, our results still indicate that there is increased airway instability in antigen-challenged and -sensitized mice, whether in large or small airways.
There is reason to believe that the inflamed airways of the antigen-sensitized and -challenged mice become unstable. Airway inflammation results in leakage of plasma proteins onto the airway surface (49), and plasma proteins are known to increase airway surface tension, thereby decreasing airway stability (10, 16). Furthermore, we have shown previously that the formation of extravascular fibrin on the luminal surface of the lung can induce AHR with a prominent H component similar to what we observed in antigen-sensitized and -challenged mice (46).
In addition to the difference we observed between the two mouse models, there were differences between the response to aerosolized and the response to intravenous methacholine. Intravenous methacholine resulted in a greater RN response in both models and a minimal response in H, and this is best explained by the fact that there is no systemic circulation in the peripheral airways of mice (32). Close examination of the time course of bronchoconstriction also reveals that there was less of a plateau in the RN response in the antigen-sensitized and -challenged mice when intravenous methacholine is administered (Fig. 4). This difference could represent the consequence of methacholine-induced mucus secretion and consequent luminal narrowing, or it could be that the smooth muscle response to aerosolized methacholine is different from that of intravenous methacholine. Perhaps aerosolized methacholine is not cleared as quickly as intravenous methacholine, or perhaps the airway epithelium has an effect on how aerosolized methacholine interacts with the airway smooth muscle. Such an aerosol-epithelial-airway smooth muscle interaction has been previously shown in AHR induced by cationic proteins (3, 5). However, it is important to note that when a deep inhalation is given at the end of the time course this residual elevation in RN is reversed. Deep inhalations are thought to relax airway smooth muscle and thereby reduce AHR (38). Residual tone and stretching and subsequent relaxation of the airway smooth muscle could explain this reversible plateau we see in the antigen-sensitized and -challenged mice, but if this were the cause it would be difficult to explain why there was not a similar plateau observed when intravenous methacholine was administered. A more parsimonious explanation is that the plateau is the result of mucus secretion that is stimulated more by aerosolized than by intravenous methacholine. Antigen-sensitized and -challenged mice have excess mucus production (36, 43), and it has been shown that mucus secretion is stimulated by aerosolized methacholine (37). Increased mucus in the airway lumen could result in effective airway closure and likely contributes to the rise in H observed in the antigen-challenged and -sensitized mice following methacholine administration, which would be reversible by deep inhalation.
Is there a preferred mouse model that should be used to study the pathophysiology of asthma? It is obvious that the response to methacholine involves smooth muscle contraction; nonetheless, exaggerated airway narrowing is not the only possible physiological derangement that can contribute to what is measured as AHR. Physiological and recent imaging studies demonstrate that functional airway closure is not only present in asthmatics, it is inducible by methacholine, and reversible by bronchodilators (18, 24, 25, 35, 44, 47, 48). It has recently been shown, using a computational modeling approach, that the heterogeneous ventilation defects observed in asthmatic patients can be attributed to narrowing and subsequent closure of small airways (41). This suggests that physiology in antigen-sensitized and -challenged BALB/c mice more closely mimics that seen in asthmatic individuals. As such, other mouse models, like A/J mice, that have more of an exaggerated central airway response to methacholine seem less relevant. However, the phenotype of asthma is highly variable (31), and there is evidence that some patients have more or less of a central airway response especially during an exacerbation (18, 22). Therefore, the choice of mouse model and route of methacholine delivery should depend on the asthma phenotype of interest. Furthermore, while it is obvious that no mouse model is a perfect representation of clinical asthma, the physiological variability demonstrated in the present study suggests that mouse models may nonetheless provide an opportunity to study the various phenotypes of asthma in relative isolation.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Cojocaru, C. G. Irvin, H. C. Haverkamp, and J. H. T. Bates Computational assessment of airway wall stiffness in vivo in allergically inflamed mouse models of asthma J Appl Physiol, June 1, 2008; 104(6): 1601 - 1610. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. K. A. Lundblad, J. Thompson-Figueroa, G. B. Allen, L. Rinaldi, R. J. Norton, C. G. Irvin, and J. H. T. Bates Mucous Obstruction and Airway Hyperresponsiveness in Mice Am. J. Respir. Crit. Care Med., May 15, 2008; 177(10): 1171 - 1172. [Full Text] [PDF]
|
|
|
|
|
|
C. Pantano, J. L. Ather, J. F. Alcorn, M. E. Poynter, A. L. Brown, A. S. Guala, S. L. Beuschel, G. B. Allen, L. A. Whittaker, M. Bevelander, et al. Nuclear Factor-{kappa}B Activation in Airway Epithelium Induces Inflammation and Hyperresponsiveness Am. J. Respir. Crit. Care Med., May 1, 2008; 177(9): 959 - 969. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. T. Bates, A. Cojocaru, H. C. Haverkamp, L. M. Rinaldi, and C. G. Irvin The Synergistic Interactions of Allergic Lung Inflammation and Intratracheal Cationic Protein Am. J. Respir. Crit. Care Med., February 1, 2008; 177(3): 261 - 268. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. F. Alcorn, L. M. Rinaldi, E. F. Jaffe, M. van Loon, J. H. T. Bates, Y. M. W. Janssen-Heininger, and C. G. Irvin Transforming Growth Factor-beta1 Suppresses Airway Hyperresponsiveness in Allergic Airway Disease Am. J. Respir. Crit. Care Med., November 15, 2007; 176(10): 974 - 982. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. T. Bates Point:Counterpoint: Lung impedance measurements are/are not more useful than simpler measurements of lung function in animal models of pulmonary disease J Appl Physiol, November 1, 2007; 103(5): 1900 - 1901. [Full Text] [PDF]
|
|
|
|
|
|
J. H. T. Bates, A. Cojocaru, and L. K. A. Lundblad Bronchodilatory effect of deep inspiration on the dynamics of bronchoconstriction in mice J Appl Physiol, November 1, 2007; 103(5): 1696 - 1705. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. C. Haverkamp and L. K. A. Lundblad Commentary on "The role of the large airways on smooth muscle contraction in asthma" J Appl Physiol, October 1, 2007; 103(4): 1463 - 1463. [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Johnson, F. K. Swirski, B. U. Gajewska, R. E. Wiley, R. Fattouh, S. R. Pacitto, J. K. Wong, M. R. Stampfli, and M. Jordana Divergent immune responses to house dust mite lead to distinct structural-functional phenotypes Am J Physiol Lung Cell Mol Physiol, September 1, 2007; 293(3): L730 - L739. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. M. Bozanich, T. Z. Janosi, R. A. Collins, C. Thamrin, D. J. Turner, Z. Hantos, and P. D. Sly Methacholine responsiveness in mice from 2 to 8 wk of age J Appl Physiol, August 1, 2007; 103(2): 542 - 546. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Carey, J. W. Card, J. W. Voltz, E. R. Jacobs, A. Dakhama, G. Larsen, E. W. Gelfand, and D. C. Zeldin Airway Responsiveness Should Be a Measurement of the Responsiveness of Airways Am. J. Respir. Crit. Care Med., July 15, 2007; 176(2): 215a - 216. [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Paveglio, J. Allard, J. Mayette, L. A. Whittaker, I. Juncadella, J. Anguita, and M. E. Poynter The Tick Salivary Protein, Salp15, Inhibits the Development of Experimental Asthma J. Immunol., June 1, 2007; 178(11): 7064 - 7071. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||
plot, mice with antigen-induced AHR; 
plot, A/J mice. 
