|Home | About | Journals | Submit | Contact Us | Français|
We quantified the effects of airway wall remodeling upon airway smooth muscle (ASM) shortening. Isolated ASM from sheep was attached to a servo-controller that applied a physiologic load. This load could be altered to reflect specified changes of airway wall geometry, elasticity, parenchymal tethering, transpulmonary pressure (PL), and fluctuations in PL associated with breathing. Starting at a reference length (Lref), ASM was stimulated with acetlycholine and held at constant PL of 4 cm H2O for 2 h. When all compartments were thickened to simulate the asthmatic airway but PL was held fixed, ASM shortened much more than that in the normal airway (to 0.52 Lref versus 0.66 Lref). When breathing with deep inspirations (DIs) was initiated, within the first three DIs the ASM in the normal airway lengthened to 0.84 Lref, whereas that in the asthmatic airway remained stuck at 0.53 Lref. Thickening of the smooth muscle layer alone produced the greatest muscle shortening (to 0.47 Lref) when compared with thickening of only submucosal (to 0.67 Lref) or only adventitial (to 0.62 Lref) compartments. With increased ASM mass, the ASM failed to lengthen in response to DIs, whereas in the airway with thickened submucosal and adventitial layers ASM lengthened dramatically (to 0.83 Lref). These findings confirm the long-held conclusion that increased muscle mass is the functionally dominant derangement, but mechanisms accounting for this conclusion differ dramatically from those previously presumed. Furthermore, increased ASM mass explained both hyperresponsiveness and the failure of a DI to relax the asthmatic airway.
Asthma is characterized by airways that constrict too easily (airway hypersensitivity) and too much (airway hyperresponsiveness, AHR) (1). It is the excessive airway narrowing associated with AHR, rather than the hypersensitivity, that accounts for the morbidity and the mortality that is attributable to the disease (2–4). AHR is thought to arise as a result of ongoing and irreversible remodeling of the airway wall (2, 5–10). Among the various factors that come into play, AHR might be accounted for either by increased mass of airway smooth muscle (ASM) or by decreased load against which the ASM must contract, but it is widely believed that increased ASM mass is the main culprit (11–13). This important conclusion is based mainly upon theoretical considerations, however, wherein structural evidence has been incorporated into detailed mathematical models of airway narrowing (11, 12, 14).
While these mathematical models incorporated the best information then available, they rested upon pivotal assumptions that were shown subsequently to be erroneous (15–18). In particular, muscle was represented by its isometric force–length relationship. But being a characterization of muscle contraction in purely static circumstances, such a description is intrinsically incapable of broaching issues involving smooth muscle dynamics as occur during tidal breathing and deep inspirations (DIs). For example, as early as 1859 Salter (19) reported the following:
…[bronchial] spasm may be broken through, and the respiration for the time rendered perfectly free and easy, by taking a long, deep, full inspiration. In severe asthmatic breathing this cannot be done; but in the slight bronchial spasm that characterizes hay asthma I have frequently witnessed it. It seems as if the deep inspiration overcame and broke through the contracted state of the air-tubes, which was not immediately re-established.
Subsequently we came to learn that of all innate agencies of bronchodilation the most potent of all is a DI (20), but in the asthmatic the ability of a DI to relax airway smooth muscle and dilate the airway is greatly attenuated (21–26). Indeed, Fish and coworkers (27) went so far as to suggest that the inability of an individual with asthma to dilate the airway with a DI might be the proximal cause of excessive airway narrowing in asthma.
Why is the individual with asthma, but not the normal subject, refractory to the potent bronchodilatory effects of a DI? And, more generally, what is the role of DIs in AHR? These questions involve airway dynamics, and here we have examined these questions experimentally; our strategy departed in two ways from those previously used. First, rather than using a theoretical description of muscle behavior as described above (11, 12, 14), here we studied smooth muscle shortening in the muscle bath. Second, as in previous approaches we used a mathematical description of the load against which the muscle is contracting; the load was dynamic rather than static, and this dynamic load was applied to real activated muscle through the agency of a servo-controller. As closely as possible, therefore, the ASM was loaded as it would be in vivo. The advantage of this hybrid strategy is that, first, it retains control of the many pathophysiologic factors that set the muscle load and, second, it makes no assumptions about the muscle itself. In contrast with previous investigations (11, 12, 14), muscle shortening can be observed directly rather than predicted theoretically. Here, however, we have adapted the approach of Latourelle and colleagues (28) so as to incorporate progressive remodeling of the different airway wall compartments.
As in many previous studies of AHR, results presented here again point to muscle mass as being the main culprit. However, these results show that mechanisms accounting for AHR are mainly dynamic and are therefore unaccounted for in previous theoretical analyses, which are static. Moreover, these results show that increased ASM mass not only explains AHR but also accounts for the failure of a DI to relax the asthmatic airway, much as had been described by Salter (19). Indeed, these experiments imply that the failure of a DI to relax the asthmatic airway is the proximal cause of AHR.
We followed the approach of Latourelle and coworkers (28). In brief, airway smooth muscle was subjected to a virtual load created by a servo-controlled lever system (Cambridge Technologies, Lexington, MA, currently sold by Aurora Scientific, Aurora, Canada). The servo-controller was programmed using a mathematical model that closely approximates the in vivo dynamic loading conditions that are believed to prevail during normal spontaneous breathing. The mathematical model took into account the following factors: (1) airway wall geometry, (2) passive airway wall forces, (3) parenchymal interdependence, and (4) transpulmonary pressure, and its changes in time. For any given muscle strip in the bath, actual muscle mass was of course fixed, but we could simulate the effects of virtual increases in muscle mass by proportionate decreases in the servo-controlled real muscle load. The passive structures in the wall were modeled using pressure–area relationships described by Lambert and colleagues (29), which were based on experimental data from Hyatt and coworkers (30). We used their parameters for airway generation 8. Constants used for calculating forces due to the passive structures of the airway wall were as follows: αo = 0.213, αo′ = 0.174, n1 = 1, n2 = 10, Am = 10.2 cm2 (29). The contribution of parenchymal distortion to muscle load was calculated using the shear modulus of lung parenchyma, μ, and the change in adventitial diameter (31); distortion of lung parenchyma imposes an increased load on the airway smooth muscle as it shortens. The shear modulus depends directly upon transpulmonary pressure, PL, according to the relationship μ = 0.7 PL(32). Transpulmonary pressure, smooth muscle force, and airway radius were directly related using the Laplace relationship in a thin-walled cylinder. Changes in adventitial diameter were related to changes in muscle length as described by Lambert and colleagues (12). These formulations of the load characteristics have already been described in detail (28). To vary airway wall geometry we used the relationships between airway size and different airway wall components described by Lambert and coworkers (12). The airway wall was subdivided into submucosal, smooth muscle, and adventitial wall areas and incorporated into the model. We could then control the thickness of each of these individual compartments to simulate remodeling of the asthmatic airway. We used the slope and intercept values for the regression of wall areas on basement membrane perimeter provided by Lambert and colleagues (12) for these calculations.
The sheep tracheal smooth muscle strip was mounted in a muscle bath (Krebs-Henseleit solution, 37°C, aerated with 95% O2–5% CO2, pH 7.3–7.5) and set to a reference length (Lref) for optimal response in the standard manner using electric field stimulation (EFS). The reference length Lref was set as the length at which the response to an EFS, 5 min after a 0.5-mm length increase, differed by < 10% when compared with the response at the previous length. After 45 min at Lref, the muscle was activated with acetylcholine chloride (ACh) and allowed to contract isometrically for 30 min. The concentration of ACh was maintained at 10−4 M by constantly pumping fresh solution in to muscle bath. The force generated by the muscle at the end of the 30 min of isometric contraction was denoted as Fref. ACh was then washed out and the muscle returned to a relaxed state.
In designing our experimental protocol, tidal breathing might have been imposed either before or after initiating the contractile stimulus. The former choice is certainly physiologic but would not allow us to obtain the appropriate control state for this particular study, which was designed to compare muscle lengths equilibrated in statically loaded versus dynamically loaded circumstances. We opted, therefore, for the latter choice, with which we were able to create a well-defined statically equilibrated control state upon which tidal loading could then be superposed. In addition, we could use each strip as its own control.
We fixed PL at 4 cm H2O to simulate the load during static conditions, activated the muscle with ACh (maintained at 10−4 M), and allowed the muscle to shorten for 120 min to its static equilibrium length (LSE). We then simulated the effect of quiet breathing with occasional deep inspirations by imposing sinusoidal fluctuations of PL (amplitude of 1.25 cm H2O) at 12 breaths/min, with a single deep breath (a positive half sine of amplitude ΔPL = 10 cm H2O) every 6 min.
We used three steps to normalize the experimental data. (1) We first normalized the load characteristics in the following manner. For any given PL, smooth muscle length was plotted as a fraction of Lo, which was defined as the radius of the relaxed smooth muscle (i.e., when active force is zero) at a PL of 10 cm H2O. The force was normalized to Fo, which was defined as the maximum force the muscle can generate. Fo was computed using the thickness of the smooth muscle at Lo and an estimated maximal stress the smooth muscle is capable of generating (we used 150 kPa ). (2) We then scaled the experimental Lref and Fref to Lo and Fo. (3) Finally, we normalized the experimental force and length to the re-scaled Lref and Fref.
We used only one level of agonist concentration: ACh at 10−4 M. Graded levels of muscle activation, mass, and different force generation capacities of the muscle were simulated by the servo-controller, which used an activation factor. The activation factor scaled the load in proportion to the Fref of the actual muscle that is in the bath. However, muscle edema or swelling, rather than variations in muscle accumulation, could not be addressed in our protocol.
To study the contributions of the different compartments of airway wall we measured changes in muscle length after replacing one of each of the compartments with modified values. In this manner we studied the effect of thickening the smooth muscle layer, the adventitial layer, or the submucosal layer. Similar to Lambert et al. (12) we did not move the luminal boundary inward while increasing the submucosal thickness.
DI amplitude was increased from 0 to 30 cm H2O in steps of 5 cm H2O every 30 min.
Different protocols used different numbers of strips (n). All data were expressed as mean ± SE. To evaluate the differences between means, t tests and ANOVA were performed using EXCEL.
The load characteristic is the elastic loading force, F, against which the smooth muscle would have to contract to attain a given muscle length, L. Being a property of the load and not the muscle, the load characteristic is determined by airway geometry, airway wall properties, tethering of the airway to the lung parenchyma, and the PL. As has been shown by others previously (12, 14, 33), load characteristics for normal airway (Figure 1, dark lines) and asthmatic airway (Figure 1, light lines) show dramatic differences.
We begin by considering only one of these load characteristics, the case of the normal airway with PL held fixed at 4 cm H2O. If muscle force is zero then the muscle length will be close to 0.9 Lo. As muscle force increases, muscle length decreases, following a sigmoidal relationship until, at ~ 0.3 Lo, muscle shortening becomes limited by airway closure. Increasing PL causes the relationship to shift upward and to the right, indicating that greater muscle force would be required to attain any given muscle length, but because airway closure is set by mainly geometrical factors, it is insensitive to PL. Moreover, airway closure occurs while F remains less than Fo, indicating that even in the normal airway, activated muscle has the force-generating capacity to close the airway (34).
In the asthmatic airway each of the airway compartments are thickened, and as a result corresponding geometric factors are altered and the load characteristics are shifted down and to the right. For the same PL = 4 cm H2O, if muscle force is zero then muscle length is about the same as in the normal airway, but as muscle force increases muscle length decreases far more than in the normal lung. Moreover, muscle shortening becomes limited at a substantially greater muscle length, and that limitation occurs at a dramatically smaller muscle force.
Using the servo-controller, we loaded airway smooth muscle with the characteristic shown in Figure 1 for the case of PL held fixed at 4 cm H2O. We then activated the muscle maximally using 10−4 M ACh. Such a contraction is neither isometric nor isotonic, but rather is auxotonic and much closer to the realistic physiology. The trajectory of muscle shortening (Figure 2A, and on an expanded scale in Figure 2C) begins at point A and follows the load characteristic until shortening eventually stops at point B, at which point the elastic load is balanced by the steady-state active force generated by the muscle. We allowed 120 min for muscle shortening to be completed and the system to come to this balance of static forces. Compared with the normal airway, in the asthmatic airway the smooth muscle shortened far more (Figures 2B and 2D).
After 120 min of auxotonic shortening against a fixed load characteristic (point B, Figure 2C), we initiated changes of transpulmonary pressure in time mimicking the pattern of normal respiration, namely, tidal breathing punctuated by occasional DIs (35).
These changes of PL caused the load characteristic to change smoothly in time between isopleths shown in Figure 1. The response to a DI was as follows. As PL reached a maximum of 14 cm H2O, the trajectory (B to C) shows that force grew substantially and exceeded Fref for much of the time and muscle length increased appreciably. As PL returned to 4 cm H2O, the muscle shortened from C back to point D′, and over the next few tidal breathing it shortened further (from D′ to D). However, as tidal breathing continued the muscle re-lengthened (in a looping trajectory) and the mean muscle force slowly decreased (D to E, blue). After 6 min, this slow lengthening was interrupted by the second DI, whereupon the muscle followed the trajectory EFG (pink). As tidal breathing and DIs continued, the muscle became equilibrated on a dynamic trajectory approximating G. At all points on this dynamic trajectory, muscle length far exceeded the muscle length at the statically equilibrated value indicated by point B.
To better visualize muscle shortening in response to this loading pattern, we calculated the average muscle length over the duration of a single breath and plotted it against time. Results from a representative muscle are shown in Figure 3 (points B to G are marked for reference to Figure 2). After 2 h of static loading, the muscle in the asthmatic airway had shortened much more than the muscle in the normal airway (P < 0.05). In the normal airway, the first DI caused muscle to lengthen substantially and had a long-lasting effect; muscle length became dynamically equilibrated. In contrast, muscle in asthmatic airway was shorter and refractory to the effects of DIs. Each DI caused only a small increase in muscle length, the effect was short-lived, and after the DI was completed the muscle quickly returned to its static frozen state.
In the normal airway (Figures 2C and and3,3, dark line), DIs and tidal breathing are thus seen to exert potent bronchodilatory effects. The first DI caused the muscle to lengthen by ~ 20%, and continued tidal breathing and DIs resulted in more lengthening. After 30 min of breathing and DI, the muscle lengthened from its static value of 0.66 ± 0.04 to 0.84 ± 0.01 (P < 0.01). By contrast, in the asthmatic airway (Figures 2D and and3,3, light line) the effects of DIs and tidal breathing were dramatically attenuated. The first DI caused the muscle to lengthen only by 5%, and by the time of the next deep inspiration 6 min later, the muscle had shortened almost back to where it was under static conditions. The muscle in this case made comparatively smaller force length excursions and appeared to be stuck at a small length (static 0.52 ± 0.05 and 0.53 ± 0.05 after 30 min of breathing with DI).
During dynamic loading, differences in muscle length between the asthmatic and normal airway became progressively smaller as muscle activation was decreased (Figure 4). However, in the normal airway changes of muscle length with change of activation were slight, whereas in the asthmatic airway they were substantial. Error bars denote SE between muscle strips and asterisk indicates a significant difference between the normal and asthmatic conditions (P < 0.05) at each level of activation. The dagger symbol indicates that for the normal muscle, there was a significant difference between its static equilibrium length and the length after 30 min of breathing and DI (P < 0.01). Importantly, the asthmatic loading condition (relative to normal) reflects not only AHR (excessive narrowing) but also hypersensitivity (substantial narrowing at low levels of muscle activation).
Against a steady load characteristic, muscle shortening increased progressively as muscle contractility increased (Figure 5). However, against a dynamic load, muscle shortening changed little up to a 2-fold increase in contractility. For greater increases of contractility and 100% muscle activation, muscle shortened to static levels.
During static loading, the statically equilibrated muscle length was sensitive to the thickness of the wall layers (Figure 6). However, when compared with thickening of only submucosal (to 0.67Lref) or only adventitial (to 0.62Lref) compartments, thickening of the smooth muscle layer alone produced the greatest muscle shortening (to 0.47Lref) (P < 0.01). During dynamic loading, the airway with thickened submucosal and adventitial layers ASM lengthened dramatically (to 0.83 Lref) (P < 0.01), but DIs and tidal breathing failed to lengthen ASM with increased ASM mass. Only when the thickness of the smooth muscle layer decreased down to normal values did the muscle re-lengthen. Of these three airway compartments, therefore, by far the most important contributor was the thickness of the muscle compartment.
When DIs were included, the normal airway equilibrated at muscle lengths that were greater as the amplitude of the DIs increased, but amplitudes in excess of 20 cm H2O did not cause any more relaxation (Figure 7). In the asthmatic airway, by contrast, the muscle remained virtually stuck at its statically equilibrated lengths even after DIs were introduced (P < 0.01), and lengthened only in response to DI amplitudes of 20 cm H2O or more.
In this report we focused upon the muscle load, the muscle mass, and the implications of changes in these features in airway narrowing in asthma. During imposed tidal breathing and DIs, but not during static loading conditions, altered airway wall geometry caused ASM to recapitulate cardinal features of asthma, including AHR, airway hypersensitivity, and the failure of a deep inspiration to reverse airway narrowing. Within the first three DIs, maximally activated ASM in the normal airway lengthened appreciably, whereas that in the asthmatic airway remained stuck at a small muscle length. When compared with the effects of thickening of submucosal or adventitial compartments, it was thickening of the smooth muscle layer that produced the greatest ASM shortening.
These findings reinforce the long-held conclusion that the functionally dominant derangement accounting for AHR is increased smooth muscle mass. In doing so, however, these findings establish that the main mechanisms underlying AHR are intrinsically dynamic, and are therefore unaccounted for in previous theoretical models, which are intrinsically static. Moreover, these findings replicate not only the potent bronchodilatory effects of DIs but also the failure of DIs to dilate the asthmatic airway, much as had been described by Salter (19) more than 150 years ago but had remained since that time unexplained. Indeed, these experiments imply that the failure of a DI to relax the asthmatic airway is the proximal cause of AHR.
In the remainder of this section we highlight reasons that previous static analyses fail to provide an adequate description of AHR, and then go on to describe the essential dynamics that are now seen to control airway narrowing.
It had been widely assumed that during bronchospasm the airway continues to narrow until the active force generated by ASM comes into balance with the passive elastic load against which the muscle is contracting (11, 12, 14); the ASM had been assumed to be statically equilibrated. But since those models were published we have subsequently learned that this assumption is not even approximately correct (16, 36–38). To represent force generated by ASM, mathematical models that have been used previously begin with the classical static force–length curve. However, spontaneous breathing causes tidal lung expansion, and tidal lung expansion causes tidal stretches of airway smooth muscle. Even though they are only a few percent of muscle length, these small stretches are transmitted to the myosin head and are large enough to cause it to detach from actin much sooner than it would have during an isometric contraction (15, 16, 39–41). Thus, these small tidal stretches perturb the binding of myosin to actin, dramatically reduce the myosin duty cycle, and thereby cause active forces generated by the muscle to be far smaller than they would be during an isometric contraction with the same degree of activation and the same mean muscle length. As a result, instantaneous muscle force is not constrained to lie upon the classical static force–length curve or the classical Hill force–velocity curve. Even when the mean muscle load is the same, differences in muscle response to static versus time varying loading conditions are rather dramatic, as illustrated in Figure 3.
Moreover, we have also come to learn that ASM can adapt its static force–length curve rather dramatically, and can do so on time scales as short as hours, minutes, or even seconds (37, 38, 42). The mechanisms of cytoskeletal remodeling that account for that adaptation are not well understood and are the subject of much current attention (42, 43). It is clear, however, that the classical force–length curve and its associated optimal length, if they exist at all, are not particularly relevant to the question of airway narrowing in asthma. Rather, through the actions of cytoskeletal remodeling the airway smooth muscle cell has the capacity to generate the same high static force over a huge range of muscle lengths, and for that reason the classical force–length curve, the optimal length, and the models that rest upon them engender errors that are not small.
Before 1992 the picture of airway narrowing was qualitative and predicted intuitively that greater muscle mass would lead to greater active force and, hence, greater airway narrowing (Figure 8A). Wiggs and coworkers (11) and Lambert and colleagues (12) performed the first quantitative analyses of airway narrowing (Figure 8B). Although remodeling of various airway compartments were shown to decrease the load and thus contribute importantly to the extent of airway narrowing, the dominant effect was found to be increased muscle mass.
In the mid-1990 s many investigators recognized that tidal loading is a potent inhibitor of active muscle force (15, 20, 36, 44, 45) (Figure 8C). While this picture helped to explain why tidal breathing and deep inspirations are potent bronchodilators, it failed to explain why individuals with asthma are refractory to the beneficial effects of a DI. Finally, it was recognized that if muscle mass becomes too large or imposed force fluctuations become too small, then ASM stiffens and therefore stretches less, and because it stretches less it stiffens even more (15, 16, 39, 40, 46), and so on. This vicious positive feedback (Figure 8D) causes the system to collapse to a statically equilibrated state; the ASM virtually freezes in the latch state, and becomes so stiff that tidal breathing and DIs can no longer perturb myosin binding.
Indeed, we have recently come to learn that the contracted airway smooth muscle cell becomes refractory to a DI because its cytoskeleton fails to fluidize, and instead remains frozen in a stiff solid-like phase (47, 48).
In conclusion, we have shown here that increased ASM mass has the potential to dominate the mechanical response of the remodeled asthmatic airway, and that ASM dynamics play a crucial role in that process. Nonetheless, several questions remain open. Neural or humoral mechanisms associated with DIs cannot be ruled out, and we did not address here differences in the muscle cell itself. For example, An and coworkers have recently studied ASM cells obtained from the relatively hyperresponsive Fisher rat and the relatively hyporesponsive Lewis rat (49). They found that differences in muscle contractility and other physical differences measured at the level of the single cell in vitro were consistent with strain-related differences in airway responsiveness in vivo. These observations highlight the fact that it remains unclear if muscle mass and muscle contractility might co-vary. It is possible, moreover, that the ASM cell in the proliferative/synthetic/maturational state might be less contractile than similar cells differentiated into fully the contractile state—an effect that would be compensatory—but no mechanical data are yet available to support that possibility. The effects of time varying mechanical loads upon ASM proliferation, phenotype, or matrix deposition, composition, and mechanical properties remain unclear as well.
This work was supported by National Heart, Lung, and Blood Institute grants HL-59682 and HL-33009 to J.J.F.
Originally Published in Press as DOI: 10.1165/rcmb.2006-0418OC on April 26, 2007
Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript.