Human airway smooth muscle is structurally and mechanically similar to that of other species

Airway smooth muscle (ASM) plays a vital role in the exaggerated airway narrowing seen in asthma. However, whether asthmatic ASM is mechanically different from nonasthmatic ASM is unclear. Much of our current understanding about ASM mechanics comes from measurements made in other species. Limited data on human ASM mechanics prevents proper comparisons between healthy and asthmatic tissues, as well as human and animal tissues. In the current study, we sought to define the mechanical properties of healthy human ASM using tissue from intact lungs and compare these properties to measurements in other species. The mechanical properties measured included: maximal stress generation, force–length properties, the ability of the muscle to undergo length adaptation, the ability of the muscle to recover from an oscillatory strain, shortening velocity and maximal shortening. The ultrastructure of the cells was also examined. Healthy human ASM was found to be mechanically and ultrastructurally similar to that of other species. It is capable of undergoing length adaptation and responds to mechanical perturbation like ASM from other species. Force generation, shortening capacity and velocity were all similar to other mammalian ASM. These results suggest that human ASM shares similar contractile mechanisms with other animal species and provides an important dataset for comparisons with animal models of disease and asthmatic ASM.

W hile the physiological function of airway smooth muscle (ASM) is unclear [1][2][3], its contraction and subsequent shortening act to increase airway resistance in the lung. Asthma is characterised by episodes of increased airway resistance due to exaggerated airway narrowing. Although it is generally agreed that the narrowing is caused by ASM shortening, it is still unclear if the excessive narrowing is due to fundamental changes in the phenotype of the smooth muscle itself, or if it is caused by structural and/or mechanical changes in the noncontractile elements of the airway wall, or by alterations in the relationship of the airway wall to the surrounding lung parenchyma [4]. One major hurdle in determining whether ASM is truly dysfunctional in asthma is the lack of adequate data on the mechanical properties of human ASM. Much of our current knowledge about ASM contractile function has come from studies in other mammalian species, which may not adequately represent human tissue. Hence, in order to investigate a potential role for ASM in asthma, it is first necessary to characterise the mechanical properties of nonasthmatic ASM. Since many hypotheses concerning the pathogenesis of asthma have been developed from animal models, it is necessary to determine the relevance of these animal models with respect to human tissues. Thus, the goal of this study was to describe the mechanical properties and ultrastructure of nonasthmatic ASM and compare the data to previous human and mammalian studies of this tissue.
Besides measuring the ''classical'' force-length and forcevelocity relationships, we have focused on a newly recognised property of ASM, namely its plastic adaptability to the dynamic lung environment. None of the previous studies on mechanical properties of human ASM have studied plastic adaptation of the tissue. Importantly, in this study we used the in situ muscle length as the reference length for the muscle instead of the previous convention of L max (the length at which force was maximal at a given instant) [23].

Tissue preparation and equilibration
Tracheas were removed from nontransplantable human lungs donated for research through the International Institute for the Advancement of Medicine (Edison, NJ, USA). All donors had no known respiratory disease and their deaths were sudden. The subject demographics and clinical details are shown in table 1. The study was approved by the University of British Columbia/St Paul's Hospital ethics committee. The whole lungs were obtained as previously described [24]. Briefly, after surgical removal the lungs were flushed with Custodiol HTK solution (Odyssey Pharmaceuticals, East Hanover, NJ, USA) and transported by plane on ice. The average time between harvesting and arrival at the University of British Columbia was 15-20 h.
The procedure of dissection and preparation of the smooth muscle was the standard procedure for sheep tissue and has been previously described, for details see [25,26]. Briefly, the tracheal tissue was kept at 4uC in physiological saline solution (PSS). Dissection was undertaken within a day of obtaining the tissue. The in situ length of the tracheal smooth muscle was used as a reference length (L ref ), which was determined prior to cutting open the C-shaped cartilage ring. Connective tissue and epithelium was carefully removed to isolate a smooth muscle bundle. Muscle strips measuring 1-1.5 mm wide, 0.5 mm thick and 6 mm long, were attached on both ends with aluminum foil clips and mounted vertically on a forcelength transducer. Before beginning the experimental protocol, the muscle was equilibrated at L ref by periodic electrical field stimulation (EFS) at 5-min intervals for a period of 1.5 h. Since some tissues exhibited substantial leukotriene-mediated tone without extrinsic stimulation, the CysLT1 receptor antagonist montelukast (10 -6 M) was added to the PSS for all of the experiments, which prevented or eliminated tone.

Mechanical measurements
After equilibration, the maximal isometric force produced in response to EFS at L ref was determined (herein called F max ). For the force-length relationship the muscle was either stretched or shortened in a stepwise fashion as previously described [25,27] and allowed to adapt to every new length for 20 min, during which it was stimulated with EFS at 5-min intervals. Five different lengths were examined: 0.5, 0.75, 1.0, 1.25, and 1.5 L ref . To determine the response to length perturbation, a 10 min, 0.2 Hz, 30% L ref length oscillation was applied. The recovery of EFS-induced (active) force was followed for 30 min after oscillation by stimulating with EFS at 5-min intervals. For the force-velocity curves, velocity measurements were made after release (quick switch from isometric to isotonic contraction) at five graded loads (between zero and F max ) at L ref , as previously described [28]. Velocities were measured at two time points: one at the peak of tetanic force and one mid-way to the peak. This is due to the nonlinear nature of shortening velocity, which peaks early on as developed force reaches ,50% of F max , and settles to a lower plateau after the force reaches F max [29]. Shortening velocity against a given load was recorded 100 ms after the release, during the steady phase of shortening. The curves were fit using Hill's hyperbolic equation [30]. Maximal isotonic shortening was established by allowing a muscle to contract against a preload equal to 10% and 20% F max . It was not possible to perform the shortening at zero load, because the servo-system of the myograph became unstable at low loads. Therefore, the maximal shortening was extrapolated to zero load from these two points.

Histology and electron microscopy
At the end of the mechanical protocols the tissue preparations were fixed at L ref in 10% formalin for histology or a formaldehyde cocktail for transmission electron microscopy (EM). The histology was performed as previously described [31]. The amount of muscle in the preparation was determined by staining transverse sections with Masson's trichrome and quantified by manual tracing using Image ProPlus 4.5 (MediaCybernetics, Bethesda, MD, USA). Maximal stress generating capacity was determined by dividing F max (in mN) by the cross-sectional area (in mm 2 ) of muscle present in the preparation. The protocol for EM followed the standard procedure previously described in our lab [32].

Statistical analysis
Force and length measurements were normalised to F max or L ref , respectively, and expressed as fractions of F max and L ref .
Velocity of shortening was expressed as L ref ?s -1 . Aggregate data were expressed as mean¡SEM. ANOVA and regression analyses were accomplished using GraphPad Prism 5 (GraphPad Software, Inc., La Jolla, CA, USA). pf0.05 was considered to be sufficient to reject the null hypothesis.

Muscle bundle properties and morphology
Human lungs were received from six donors with an average age of 25.8¡2.8 yrs. After measurements of mechanical properties, muscle strips were fixed and stained with Masson's Trichrome for histological morphometry (n54). From the histological sections, the cross-sectional area of the muscle was measured and compared to the entire tissue preparation ( fig. 1a). The percentage of smooth muscle crosssectional area to the total cross-sectional area of the tissue preparation averaged 36.5¡0.04%. The maximal stress generated by the muscle averaged 82.1¡17.3 mN?mm -2 .

Ultrastructure
Muscle strips were also fixed for EM. Areas of electron dense material, known as dense bodies and plaques, are also seen. Mitochondria, sarcoplasmic reticula, caveolae and microtubules are also present and are indistinguishable from their counterparts in animal cells.

Force-length properties and length adaptation
The relationship between muscle length and force-generating capacity was examined by recording force at five different lengths (  connective tissue blue, and nuclei black. The yellow outline encircles the muscle bundles. The glandular, red staining cells were excluded. Morphometric analysis was used to determine the cross-sectional area of muscle in the dissected strip. This area was used to calculate the maximal stress generated by each smooth muscle preparation (stress is equal to force divided by cross-sectional area). Colour segmentation was used to exclude the blue connective tissue within muscle bundles. Image width 1.14 mm. adaptation (p,0.05 and p,0.001, respectively). Passive force (figs 2c and 3) also demonstrated a significant length adaptation (p50.0053, ANOVA). At lengths longer than L ref , the passive force initially increased with the length change but decreased over the period of adaptation. Conversely, at shorter lengths, the passive force initially declined but gradually increased over the 20-min period. On average, the passive force did not reach pre-length change levels, despite significant adaptation ( fig. 3).

Recovery from mechanical perturbation
The magnitude and time required for force recovery following a perturbation was examined by applying an oscillatory strain to the relaxed muscle. The 10 min length oscillation occurred at a frequency that mimics human breathing (0.2 Hz) and at a magnitude similar to deep inspiration (30% L ref ). As seen in figure 4 (not corrected for force deterioration), immediately following oscillation the force decreased to 0.643¡0.03 F max and recovered to 0.953¡0.03 F max 30 min after oscillation.

Shortening velocity
Two sets of data were generated to examine the shortening velocity of human ASM. At both time points, the muscle was released to five different pre-determined loads (each representing a certain percentage of F max ). The difference between the two sets of data was the time of release ( fig. 5). The late releases were performed during the tetanic plateau of active force, while the early releases were performed mid-way to the plateau (on average 3-4 s after the onset of stimulation

Maximal isotonic shortening
In four muscle preparations the extent of unloaded shortening was determined ( fig. 7). Each ASM strip was allowed to isotonically shorten against a predetermined load (10% and 20% F max ). The maximal unloaded shortening was extrapolated from a linear regression of the points (R 2 50.749, p50.0055). This regression assumes that the ascending limb of the force-length curve for a nonadapted muscle is linear [33]. The 95% confidence intervals ranged from 60.3 to 84.1% of shortening at zero load with an average of 72.2% (total length of the muscle).

DISCUSSION
Despite the prominent role of ASM in airway diseases, the basic mechanical properties of human ASM have not been adequately determined. This study is the most comprehensive description of the mechanical properties of human ASM to date and is unique in that the experiments were performed within the paradigm of ASM mechanical plasticity using a high-quality tissue source. Our results demonstrate that human ASM possesses similar mechanical properties and morphological features as those found in other mammals such as dogs, pigs and sheep. This provides justification for using ASM tissues from non-human species to elucidate contractile mechanisms. Also, mechanical responses to experimental interventions in non-human mammalian ASM can now be justifiably interpreted in terms of human airway physiology. This study also provides a database of mechanical parameters for intact nonasthmatic human ASM, an invaluable reference against which asthmatic ASM can be compared.
Qualitatively, human ASM is indistinguishable from that of other species, as concluded from comparisons with ultrastructural images [25,[33][34][35]. The present study demonstrates the  feasibility of quantitative comparisons of ultrastructural features (e.g. filament densities) and the relative abundance of mitochondria, caveolae, sarcoplasmic reticula, and other organelles of interest. Comparisons of these features will provide invaluable insights into the phenotypic changes associated with asthmatic ASM.
The results of previous studies suggested that human ASM possesses reduced stress generating capacity, reduced shortening capacity and greater passive stiffness compared with other species [5,6]. Human ASM was found to generate a maximal stress of 50¡20 mN?mm -2 compared with 140, 72 and 80 mN?mm -2 in rabbit, dog and swine ASM, respectively [6]. The present results show that human ASM (82¡17.3 mN?mm -2 ) generates stress that is comparable to dog and swine ASM. Another major finding of the present study is the ability of human ASM to undergo extensive shortening (72%) ( fig. 7). This degree of shortening is similar to the reported values in other animal species, which range from 61 to 71%, and is much greater than the previously reported human ASM value of 25¡9.0% [6]. This discrepancy could be related to the reduced stress generated by the tissues in earlier studies which were dissected from bronchi obtained from surgical resections. In addition, the starting length of the muscle was set at L max , an arbitrary length that is greatly influenced by the state of adaptation of the muscle [23].
Another difference in the present study, which could influence the degree of maximal shortening, was the percentage of muscle to total tissue area (36.49¡0.04%), which is much greater than the 8.7% reported in preparations from earlier human studies [5]. These data suggest that the previously studied human ASM was surrounded by a greater proportion of connective tissue, resulting in higher passive stiffness when compared with other species. Increased connective tissue could also prevent shortening by acting as a radial constraint and/or compressive load in parallel to the muscle cells [36]. This hypothesis is supported by the observation that treating human ASM preparations with collagenase led to a 50% increase in maximal isotonic shortening [11,12]. Our carefully dissected tracheal preparations had similar ASM content to other non-human preparations that ranged from 25% to 35% smooth muscle area [6]. While the decreased connective tissue surrounding the muscle may have increased shortening in our study, it is unclear if passive tension was affected. In the previous study conducted at L max , the passive tension in the muscle preparation was 60¡8.8% of maximal force [5]. In the present study average passive tension at L ref was 22.4¡0.1% F max .
Shortening velocity in smooth muscle peaks early in contraction, before force reaches a plateau; the velocity then decreases to a lower level after force plateaus and this velocity is maintained during the sustained phase of contraction. This decrease in velocity has been attributed to the development of latch-bridges caused by dephosphorylation of the myosin regulatory light chain in arterial muscle [37]. In non-human ASM, it has been proposed that the decreased velocity is likely due to thick filament lengthening during force development [28] and myosin phosphorylation was not found to correlate to velocity [38]. The present finding that velocity declines ( fig. 6) during the time course of an isometric contraction suggests that myosin filaments in human ASM undergo similar rearrangements.
This is the first study to investigate whether length adaptation occurs in human ASM. Smooth muscle, like striated muscle, operates over a length range. In striated muscle, the force-length relationship is characterised by a well-defined optimal length where developed force is at its maximum.
Outside the narrow plateau of maximal force, the force developed by the muscle decreases at shorter and longer lengths. This relationship depends on the overlap of the actin and myosin filaments within the muscle. Likewise, smooth muscle displays a force-length relationship that approximately resembles a concave-down parabolic curve. However, ASM producing ''suboptimal'' force at a shorter or longer length can adapt to the new length, increasing its force within minutes [27]. This requires rearrangement of contractile and cytoskeletal proteins, and effectively broadens the force-length relationship of the smooth muscle [35,39]. Thus, length adaptation allows smooth muscle to operate over the large length ranges seen in vivo, particularly in the smooth muscle tissues that line hollow organs that undergo large volume changes. Compared to previous studies in canine ASM [27] our data suggests that human ASM is less adaptable; that is, the length range of the force plateau in human ASM is smaller than that of canine ASM. Nevertheless, maximal force in human ASM is length-independent over the range of 0.75-1.5 L ref (p,0.05, ANOVA), a clear sign that human ASM is capable of length adaptation ( fig. 3).
The ability of the muscle to undergo length adaptation suggests that human ASM is capable of the mechanical plasticity demonstrated in other mammalian species. This ability to rearrange intracellular elements to generate force is further supported by the muscle's response to length oscillation ( fig. 4). Our laboratory has previously shown in swine tissue that an oscillatory strain causes a decrease in force immediately after oscillation, followed by an exponential force recovery that coincides with myosin filament reformation [32]. However, the magnitude of recovery is somewhat blunted in human ASM, since the force did not fully recover to F max , as it does in swine. Likewise, the rate of recovery is slower, with a rate constant of 0.139 s -1 versus 0.234 s -1 in swine ASM [32]. Human ASM is also more easily perturbed, since the force produced by human ASM immediately after oscillation was 64% F max compared with ,79% F max in swine [32]. This implies that human ASM has a more labile intracellular structure that reorganises more slowly than ASM from other species. In vivo, this may be beneficial for maintenance of airway patency. Coupled with its reduced ability to undergo length adaptation, the force loss associated with oscillation could enhance and prolong the effectiveness of bronchoprotection provided by a deep inspiration [40,41]. The increased sensitivity to mechanical perturbation could be related to differences in the muscle's contractile apparatus or due to differences in extracellular structures, perhaps exposing human ASM to greater internal forces than are seen by other species. Comparisons to ASM from asthmatic subjects are certainly needed.
Previous studies comparing asthmatic and nonasthmatic ASM mechanical properties have suggested increased contractility (force development or shortening) [12,20,22] or sensitivity to certain agonists in asthmatic ASM [12,13,17], while others have shown no differences [14-16, 18, 19, 21]. With respect to the current study, it is unclear whether these are true differences or due to the inexactness of using L max to determine ''optimal force,'' since the velocity, extent of shortening, and sensitivity to an agonist [42] are all length dependent. Only two studies have normalised force to crosssectional muscle area, concluding that asthmatics generate more stress than nonasthmatics [12,22]. Based on the results of the present study we suggest that these investigators may have underestimated maximal stress and the extent of shortening, either because of the way L max was determined or due to less than optimal tissue quality. However these factors may have equally affected asthmatics and nonasthmatics and thus would not alter their conclusions. Alternatively, it is possible that these factors could differentially affect asthmatic and nonasthmatic tissues and lead to (or conceal) differences between the two groups.
In conclusion, this study is the first comprehensive study of the mechanical properties of healthy human ASM sourced from an intact lung. We have demonstrated that human ASM has ultrastructural features and mechanical properties which are similar to other animal species and is capable of length adaptation. These results suggest that human ASM probably has the same intracellular organisation and undergoes the same processes of mechanical plasticity that have been identified in the ASM of other species.

STATEMENT OF INTEREST
A statement of interest for P.D. Paré can be found at www.erj. ersjournals.com/misc/statements.dtl