INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CA²⁺ plays a central role in activation of both smooth muscle contractility (7) and endothelium-dependent relaxation, for example via nitric oxide synthase (5). Although the exact role of epithelial factors, such as eicosanoids or nitric oxide, in modulating airway contractility is controversial (13), intracellular Ca²⁺ is likely a major factor in regulation of epithelial barrier and secretory function. Thus the mechanisms underlying Ca²⁺ homeostasis in these tissues are central to regulation of airway function. The loading, storage, and release of Ca²⁺ by the sarcoplasmic/endoplasmic reticulum are determined in part by its associated Ca²⁺-ATPases (12). These Ca²⁺ pumps are known to exist in several isoforms; in particular, the sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase 3 (SERCA3) isoform is known to be expressed in aortic endothelium and tracheal epithelial cells (2, 3). Interestingly, SERCA3 is reported not to be in tracheal smooth muscle (1). The role of this Ca²⁺ pump isoform is not known with certainty, particularly in view of the fact that SERCA2b is also known to be expressed in these epi- and endothelial tissues and has different Ca²⁺ and pH dependencies (10). To better understand the role of SERCA3, we used a recently developed mouse model in which the SERCA3 gene was ablated (9). In this mouse, we showed that endothelium-dependent relaxation in aorta was defective. Our hypothesis here is that the airway smooth muscle from the SERCA3-deficient (SERCA3) animals will be unaffected, whereas epithelial-mediated responses will be altered compared with those of wild-type tissues. To test this hypothesis, we characterized the contractility of mouse tracheal smooth muscle and epithelium-dependent relaxation of the trachea. In addition to the first direct evidence for epithelial modulation of mouse airway smooth muscle contractility, we show that SERCA3 can play a significant role in modulating airways epithelial interaction.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Tracheal tissue preparation. Heterozygous mice were mated, and the pups were reared for 8-12 wk. Age-matched wild-type and SERCA3 pairs were then selected for experimentation. Each mouse was killed by CO₂ asphyxiation. The trachea was then quickly dissected and subsequently placed in physiological salt solution (PSS) of the following composition (in mM): 118 NaCl, 4.73 KCl, 1.2 MgCl₂, 0.026 EDTA, 1.2 KH₂PO₄, 2.5 CaCl₂, and 11 glucose, buffered with 25 NaH₂CO₃ to attain a pH of 7.4 at 37°C when bubbled with a mixture of 95% O₂ and 5% CO₂. Additional fat and tissue were carefully removed from the trachea using spring scissors under a light microscope. In some experiments, the epithelium was removed by passing a silk thread or dental floss (Ultra Floss; Oral B, Redwood City, CA) through the lumen of the trachea. Details of this procedure and analysis of its effects on epithelium-dependent function are given in RESULTS.

Force measurements in tracheal ring preparations. Tracheal rings were mounted isometrically to force transducers (Harvard Apparatus, Hilliston, MA) using 0.012-in.-diameter wires such that force in the circumferential direction was measured. The wire-mounted tracheae were then immersed in a 50-ml organ bath filled with PSS and maintained at a constant temperature of 37°C. Isometric force signals were acquired using a data-acquisition program (AcqKnowledge, BioPac version 3.2). After 30-45 min of equilibration, two to three contraction/relaxation cycles to ACh (10 µM) were performed until reproducible forces were achieved.

Perfused trachea preparations. Tracheae were excised and dissected free of surrounding tissues as in Tracheal tissue preparation. Each trachea was tied with the ends around two cannulas made of 18-gauge stainless steel tubing using 5-0 silk. The span of exposed trachea between the cannulas was 5 mm. Differential pressure (P) was measured using two pressure transducers (Cobe) attached to wall taps near the cannula openings and was recorded using the BioPac data- acquisition system. Tracheae were placed in an open chamber, surrounded by 10 ml of PSS. A separate chamber of 14 ml PSS was perfused and recirculated at 16 ml/min through the lumen of the trachea using a gear pump (Micropump). At this flow rate, baseline P ranged from 1.1 to 1.5 mmHg for tracheae without stimulation.

Data analysis. Data are expressed as means ± SE. The n value indicates the number of mice used. For ring preparations, tracheae from wild-type and gene-targeted mice were mounted in the same organ bath. Standard ANOVA was used, with P < 0.05 taken as evidence of statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Tracheae were obtained from 10- to 12-wk-old mice that weighed ~32 g; there were no differences in weight between mouse types, similar to that previously reported (9). For comparison between gene-targeted and wild-type animals, sibling pairs were used. The average values of the gross morphological properties of the SERCA3 and wild-type tracheae are presented in Table 1. In terms of force-generating capacity, there were no differences in cross-sectional area, which generally is linearly related to cellular area and force capacity. There were no significant differences in physical properties of these tracheae, which is consistent with that previously reported for aortas from these mice (9).

Force-length behavior. To define the optimal length for maximum active force, we measured the force-length characteristics of the tracheal ring preparations. We recorded the passive force (resting tension) and active force (10 µM ACh) and determined the relationships from 14 tracheae, averaging eight length points each. Figure 1, A and B, shows these relationships developed from 111 experimental points. The resting tension of the trachea varied exponentially as C/C_s (Fig. 1A), where C is circumference, and C_s is the circumference of the trachea under unloaded conditions. In Fig. 1B, the active force (total force  passive force) is shown along with its SE. The force-length relationship was very broad. We chose 1.4 × C/C_s as our standard length, choosing a slightly suboptimal length to improve resolution by reducing passive force relative to the active force generated. At this length, maximum active forces per cross-sectional area averaged 8 mN/mm² and did not differ between trachea types (Table 1).

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Force-length relationships of murine tracheal smooth muscle. A: isometric force-length relationships for wild-type mice tracheae stimulated with 10 µM ACh. Top curve represents the total force of the wire-mounted tracheae. Center curve represents the tracheal resting tension, i.e., the passive force. Bottom curve represents the active ACh-induced force (total  passive force) at different muscle lengths. B: active isometric force-length relationships shown in greater resolution (solid line) and the boundaries (dotted lines) that represent the SE of the fit. In both graphs, length was normalized to C/Cs [the current circumference of trachea (C) compared with the circumference under 0 load (Cs)]. Data are from 14 tracheae and ~8 measurements at different lengths per trachea.

Concentration-active force relationships. We chose ACh for our standard contractile agonist because KCl, often used for reference contractions, produced only a small contraction in the organ bath. Figure 2 shows a recording of a typical cumulative concentration-force measurement. Thresholds were recorded at ~0.1 µM, and maximum forces occurred at 10-30 µM. Figure 3 graphically summarizes the data from tracheae from five wild-type and SERCA3 mice, which yielded an average concentration-force relationship with an ED₅₀ of 2 µM. There were no statistically significant differences in these relationships between SERCA3-deficient and wild-type tracheae.

View larger version (8K): [in this window] [in a new window]   Fig. 2.   Typical record of an experiment measuring cumulative ACh-active force relationships for mouse tracheae. Arrows indicate the point in time at which the concentration of ACh was increased (0.001, 0.003, 0.01, 0.03, 0.1, 0.3, 1, 3, 10, and 30 µM, respectively). Horizontal scale arrow represents 20 min; vertical scale arrow represents 10 mN.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   ACh concentration-active force of murine tracheal smooth muscle. Values are averages ± SE (bars) for 5 tracheae from wild-type (WT; dotted line) and sarcoplasmic/endoplasmic reticulum Ca2+-ATPase 3-deficient [(SERCA3); knockout (KO); solid line] mice. There were no statistically significant differences.

Characterization of epithelial airway smooth muscle interactions. We tested substance P (SP), ATP, and thrombin, typical endothelium-dependent vasodilators (4), for potential epithelium-dependent effects. In the organ bath, when tracheae were contracted with 10 µM ACh, SP and ATP each elicited a transient relaxation, whereas thrombin had little effect. SP (0.01 µM) and ATP (10 µM) were chosen as standards for our experiments because these concentrations maximized the peak relaxation. Due to tachyphylaxis, we did not attempt concentration-relaxation relationships. To assess whether these responses were dependent on an intact epithelium, we first tried to remove the epithelium using techniques similar to those reported for removal of vascular endothelium. We tried several techniques and were most successful with passing a thread (6-0 silk) through the lumen. Considerable care was required in this protocol in that both the magnitude of the force and relaxation were decreased by mechanical epithelial denudation. We were able to develop a protocol used on 11 tracheae in which force after passage of the silk thread was not significantly decreased (89 ± 6% of the force before denudation). In these tracheae, the peak relaxation to SP (0.01 µM) was decreased by 50 ± 4% (n = 11) compared with that before our epithelium denudation protocol. In parallel time-course studies, the controls were decreased by 12 ± 3%, significantly less than the silk thread-treated preparations (n = 5, P = 0.026). We examined the histology from one pair of tracheae in which the relaxation to SP was measured. After a control contraction and relaxation to SP, we removed the epithelium from one trachea and repeated the protocol. The tracheae were subsequently fixed in paraformaldehyde. Using three separate hematoxylin and eosin-stained sections of each trachea, epithelial cells were counted based on the number of nuclei in the epithelial layer. The control trachea retained 100% of its original relaxation response to SP, and the epithelial layer consisted of 567 ± 19 continuously adjacent cells in a cross section of the tracheal ring. The silk thread-treated trachea retained only 33% of its original SP response, and the epithelial layer contained 135 ± 8 cells evenly distributed around the ring, with no more than eight cells adjacent to each other. These results suggested that at least part of the relaxation to SP was dependent on an intact epithelium.

To further investigate whether the SP relaxation was epithelium dependent, we developed an apparatus for perfusing an isolated mouse trachea and for measuring the pressure drop due to tracheal resistance. This resistance is related to tracheal diameter and hence is a measure of tracheal contractility. This method is similar to one developed by Munakata et al. (11) for use in guinea pig trachea but was modified to accommodate the smaller dimensions of mouse airways. In this experiment, the trachea could be both perfused and superfused such that only the luminal side or exterior was exposed to the test solution. Figure 4A shows a typical experiment. After application of ACh (10 µM) to the exterior solution, a steady-state contraction was measured, as indicated by the increased pressure drop across the tracheal resistance. Application of SP to the exterior had no effect for at least 30 min (n = 5). In contrast, addition of SP to the perfusion of the luminal side of the trachea elicited a relaxation similar in magnitude and time course to that observed in the ring preparation. ATP (10 µM) applied to the external bath (Fig. 4B) produced only a small relaxation (26 ± 9%) compared with luminal administration (76 ± 3%; n = 4).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Response of murine trachea to perfused or superfused substance P (SP) or ATP. A: typical record showing that SP-induced relaxation only occurs with luminal perfusion. B: typical record showing ATP-induced relaxation is significantly greater with luminal perfusion than superfusion. C, contraction with 10 µM ACh; W, washout. Subscripts denote extraluminal (e) or intraluminal (i) administration of agents. Agents were added at the points indicated by the arrows and were not removed until the points indicated by the "washout" arrows. Horizontal scale arrow represents 20 min; vertical scale arrow represents 0.5 mmHg differential pressure.

To determine whether the differences in responses between luminal perfusion and superfusion were related to epithelial barrier or secretory function, we developed methods to remove the epithelium while the trachea remained mounted in the perfusion apparatus. To remove the epithelium from the trachea while mounted in our perfusion system, we passed a length of dental floss (Oral B Ultra Floss) through the lumen of the trachea. For three tracheae, this treatment reduced the relaxation to SP from 60 ± 6% of the steady-state contraction to 15 ± 2%. These experiments provide additional evidence for an epithelial dependence of the responses to SP and ATP.

Characterization of the SP relaxation. Pretreatment of the ring preparation for 20 min with the nitric oxide synthase inhibitor N-nitro-L-arginine (0.2 mM) decreased the relaxation induced by SP in ACh-contracted trachea by 36 ± 12% (n = 9). Pretreatment with 10 µM indomethacin increased the force of ACh contraction similarly, by 36 ± 8% in wild-type and by 38 ± 10% in knockout tracheae. Importantly, indomethacin abolished the SP-induced relaxation in all tracheae (n = 5 wild type; n = 5 SERCA3) studied. Because the extent of relaxation is often inversely related to the magnitude of contraction, it is possible that part of the decrease with indomethacin might be attributable to this factor. As a control, we studied SP relaxation using 30 mM ACh, a concentration at which force was 150% greater than our reference contraction (10 µM ACh) used for the measurements of SP effects. We found that the SP relaxation, expressed as a percent of the developed force, was independent of ACh concentration or force generated in either wild-type or SERCA3 tracheae (n = 3 each). In absolute terms, the magnitude of the peak relaxation was actually greater at 30 mM ACh, but, importantly, there were no differences between wild-type and SERCA3 tracheae. Thus the complete inhibition by indomethacin would suggest that the eicosanoid pathway is the major pathway in the epithelium-dependent, SP-induced relaxation.

Comparison of wild-type and SERCA3 tracheae. Figure 5 shows typical SP responses in SERCA3 tracheae, and averaged properties for 9-14 pairs are presented in Table 2. After ACh stimulation, SP elicited a transient relaxation that was similar in magnitude for tracheae from wild-type (42.0%, n = 14) and SERCA3 (43.1%, n = 14) mice. However, the rate of relaxation was significantly (P < 0.04, n = 9) more rapid in the wild-type [half-time (t_1/2) = 34.3 s] than in the SERCA3 (t_1/2 = 61.6 s) tracheae. To ensure simultaneous addition of SP, wild-type and SERCA3 tracheae were mounted in the same bath. It is possible that some interaction between these tissues could occur, but with the dilution associated with a 50-ml bath (a dilution factor >20,000), we feel this is unlikely. We tested for such effects by studying wild-type and SERCA3 tracheae in separate organ baths. There was no difference in the relaxation to SP from that observed when mounted in the same bath. This suggests that differences in time course were not due to diffusion of a relaxing factor from one trachea to the other. Moreover, when an epithelium-denuded and an intact trachea were placed in the same bath, the intact trachea relaxed normally to SP while the denuded trachea did not relax at all. Thus the differences in time courses do not reflect effects of diffusion of an epithelium-dependent factor.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Time course of SP-induced relaxation of SERCA3 (solid line) and wild-type (dotted line) tracheae. Representative tracings for typical SERCA3 and wild-type tracheae precontracted with 10 µM ACh and subsequently treated with 0.01 µM SP are shown. Recordings were scaled such that the magnitudes of the peak relaxations were identical. Note that the average forces per cross-sectional area did not differ between mouse types. Horizontal scale arrow represents 10 min; vertical scale arrow represents 5 mN. Wild-type trachea achieves peak relaxation well before that of the SERCA3-deficient (SERCA3) trachea.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   SP- and ATP-induced relaxation of murine tracheal smooth muscle. Representative tracings for SERCA3 (solid line) and wild-type (dotted line) tracheae precontracted with 10 µM ACh and treated with 0.01 µM SP and subsequently with 10 µM ATP are shown. Horizontal scale arrow represents 10 min; vertical scale arrow represents 5 mN. Time-to-peak relaxation in response to SP is shorter in the wild-type trachea (as in Fig. 5), whereas, in response to ATP, it is greater than that of SERCA3 trachea.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

There are few experimental data on mouse tracheal function, and our first goal was to characterize its contractility. It is likely that, with the generation of mouse transgenic and gene-ablated models, it will become an important preparation for study of airway regulation. Its force-length relationships are similar to other smooth muscles. Active isometric force per cross-sectional area was also not unusual and in the range reported for other mouse smooth muscles, e.g., aorta (14). A second objective of this study was to determine if epithelium-dependent modulation of contractility could be demonstrated. Although vascular smooth muscle-endothelial cell interactions are well characterized (4), there is less evidence for such phenomena in airways. Our evidence, based on both mechanical denudation of tracheal epithelium and side-specific delivery of agonists, indicates that the SP and ATP relaxation of an ACh contracture are epithelium-dependent responses. Although our data show a relaxation dependent on epithelium, there is no general consensus on the nature of such dependency. The current controversy centers on whether the epithelium "barrier" or "secretory" function is the major factor (13). Our data strongly support the latter, particularly in the case of SP. Application of SP to the serosal side with more direct access to the smooth muscle component of the trachea had no effect, whereas luminal infusion elicited relaxation. These results would also argue against ascribing these results to nonepithelial differences such as in the differences in localization of degradative enzymes, such as peptidases, in the case of SP, or for conversion of ATP to adenosine. Moreover, adenosine (up to 10 µM) had no effect on the ring preparations (n = 4). Thus the studies on the perfused trachea preparations make any barrier or mechanism involving differential enzyme localization unlikely to explain our results. In conjunction with the studies on the ring preparations, our data strongly suggest that the relaxation elicited by either SP or ATP is an epithelium-dependent, secretory function.

Our third objective was to identify the role of SERCA3 in epithelial function in the airway. SERCA3 belongs to the P-type superfamily of ion transport ATPases. The tissue specificity and cell-type distribution of SERCA3 is limited, in contrast to the near-universal distribution of SERCA2b. It is reported (3, 15) to be present in endothelial cells, in epithelial cells of the trachea, intestine, and salivary glands, and in platelets, mast cells, and lymphocytes. SERCA3 has a lower Ca²⁺ affinity and a higher pH optimum than SERCA2b (10). This is likely of functional significance, since SERCA2b is also found in all cells known to contain SERCA3. The limited expression of SERCA3 suggests that it may be involved in tissue-specific signaling. This is in contrast to the near-universal distribution of SERCA2b, which suggests that it likely is involved with a broad-based function in Ca²⁺ homeostasis essential for cell viability. Given the differences in biochemical parameters between SERCA3 and SERCA2b, one might also anticipate that they would operate in different cellular environments or compartments. With respect to smooth muscle function, the ACh concentration-force relationships for both wild-type tracheae and SERCA3 tracheae were similar. This is consistent with the reported distribution of SERCA3, indicating that it is not found in tracheal smooth muscle (1).

Comparison of the SP- and ATP-induced relaxation wild-type and SERCA3 tracheae suggests a role for the SERCA3 pump in modulation of epithelium-dependent processes. The major effect observed was on the rate of the relaxation. For SP, the relaxation t_1/2 was considerably shorter in the wild-type than that in the SERCA3 trachea. The relaxation was abolished by eicosanoid pathway inhibition and was partially blocked by inhibition of nitric oxide synthase. Both the eicosanoid and nitric oxide pathways, implicated by these studies, are dependent on intracellular Ca²⁺. Thus one possible mechanism to explain the slower rate of relaxation is that, in the absence of SERCA3, there is less Ca²⁺ available for release by the endoplasmic reticulum.

In response to ATP, there was a trend to the opposite direction, with the t_1/2 being the same or faster in the SERCA3 trachea than in the wild type. One might anticipate that differences in epithelial processing of SP and ATP, for example conversion of ATP to adenosine vs. SP mediator release, could underlie the differences in time courses. Adenosine did not relax these preparations, so the effects of ATP were likely mediated by purinergic receptors. A speculative but intriguing basis for this difference lies in the possibility of a differential localization of the SERCA isoforms. In tracheal epithelial cells, it is not known whether the different SERCA isoforms serve the same or different subcompartments of the endoplasmic reticulum. However, in salivary epithelial cells, SERCA2b and SERCA3 do localize to separate compartments (8). It is possible that the signaling pathway for subcompartments associated with SERCA2b may be different from that associated with SERCA3 and may underlie the differences in relaxation between wild-type and SERCA3 aortas to SP and ATP.

The physiological significance of an epithelium-dependent relaxation to SP is also not clear, but our results are not unprecedented. In organ bath studies, rubbing the epithelial layer with cotton wool blunted the SP relaxation of rat trachea (6).

In conclusion, we have shown that mouse tracheal muscle shows an epithelium-dependent relaxation involving an indomethacin-inhibitable factor. In tracheae from SERCA3 mice, smooth muscle contractility is unchanged, but the time course of the epithelium-dependent relaxation to SP is altered. Our results are consistent with earlier studies on endothelium-dependent relaxation in SERCA3 aorta. These results clearly implicate a unique role in endo(epi)thelial cell signal transduction for the SERCA3 isoform.