Airway smooth muscle relaxation results from a reduction in the frequency of Ca2+ oscillations induced by a cAMP-mediated inhibition of the IP3 receptor

Background It has been shown that the contractile state of airway smooth muscle cells (SMCs) in response to agonists is determined by the frequency of Ca2+ oscillations occurring within the SMCs. Therefore, we hypothesized that the relaxation of airway SMCs induced by agents that increase cAMP results from the down-regulation or slowing of the frequency of the Ca2+ oscillations. Methods The effects of isoproterenol (ISO), forskolin (FSK) and 8-bromo-cAMP on the relaxation and Ca2+ signaling of airway SMCs contracted with methacholine (MCh) was investigated in murine lung slices with phase-contrast and laser scanning microscopy. Results All three cAMP-elevating agents simultaneously induced a reduction in the frequency of Ca2+ oscillations within the SMCs and the relaxation of contracted airways. The decrease in the Ca2+ oscillation frequency correlated with the extent of airway relaxation and was concentration-dependent. The mechanism by which cAMP reduced the frequency of the Ca2+ oscillations was investigated. Elevated cAMP did not affect the re-filling rate of the internal Ca2+ stores after emptying by repetitive exposure to 20 mM caffeine. Neither did elevated cAMP limit the Ca2+ available to stimulate contraction because an elevation of intracellular Ca2+ concentration induced by exposure to a Ca2+ ionophore (ionomycin) or by photolysis of caged-Ca2+ did not reverse the effect of cAMP. Similar results were obtained with iberiotoxin, a blocker of Ca2+-activated K+ channels, which would be expected to increase Ca2+ influx and contraction. By contrast, the photolysis of caged-IP3 in the presence of agonist, to further elevate the intracellular IP3 concentration, reversed the slowing of the frequency of the Ca2+ oscillations and relaxation of the airway induced by FSK. This result implied that the sensitivity of the IP3R to IP3 was reduced by FSK and this was supported by the reduced ability of IP3 to release Ca2+ in SMCs in the presence of FSK. Conclusion These results indicate that the relaxant effect of cAMP-elevating agents on airway SMCs is achieved by decreasing the Ca2+ oscillation frequency by reducing internal Ca2+ release through IP3 receptors.


Introduction
A major symptom of asthma is an excessive contraction of airway smooth muscle cells (SMCs) which results in airway hyper-reactivity. To alleviate this acute and chronic airway constriction, β-adrenergic agonists, that relax SMCs, are commonly administered [1]. Yet, despite their efficacy and wide-spread application, the signaling pathways underlying the relaxing effect of β-agonists are not fully understood.
It is generally accepted that stimulation of β-adrenergic receptors activates adenylyl cyclase (AC), via receptor associated G proteins, to increase cAMP levels to mediate SMC relaxation [2]. The role of cAMP in relaxation has been confirmed with other compounds that increase intracellular cAMP concentration ([cAMP] i ), such as forskolin (FSK) that directly activates AC, theophylline that inhibits phosphodiesterase and 8-bromo-cAMP that is an analog of cAMP [3][4][5]. The traditional mechanism for cAMP action is via the stimulation of protein kinase A (PKA) to phosphorylate a variety of target proteins to induce airway SMCs relaxation. Alternatively, cAMP may act independently of PKA by interacting with exchange proteins (EPACs) [6]. EPACs have been found to activate PCLε to enhance IP 3 induced Ca 2+ release in non-smooth muscle cells [7]. However, before the specific details of cAMP-mediated signaling can be explored, it is essential to initially identify the fundamental action of cAMP on the mechanisms that regulate SMC contraction.
Airway SMC contraction is determined by the balance between phosphorylation and de-phosphorylation of the regulatory light chain of myosin (rMLC). Phosphorylation of rMLC is induced by Ca 2+ -calmodulin activated myosin light-chain kinase (MLCK). De-phosphorylation of rMLC is believed to be mediated by myosin light chain phosphatase (MLCP) [8]. As a result, SMC relaxation may be the result of different cellular pathways that culminate in either or both a reduction in MLCK activity and an increase in MLCP activity.
Perhaps the most direct method of relaxation is a reversal of the Ca 2+ response that stimulates MLCK activity and contraction. Ca 2+ signaling in airway SMCs is frequently induced by agonists such as ACh, 5-HT and ATP and consists of Ca 2+ oscillations [9][10][11]. In our recent studies with lung slices, we found that the frequency of these Ca 2+ oscillations correlated with the contractile state of the airways, a relationship that indicates the SMC tone is regulated in a frequency-modulated manner [12][13][14].
In previous studies with isolated tracheal SMCs, cAMP was found to modulate the frequency of Ca 2+ oscillations [15,16]. However, its effect on airway contraction or the mechanism of action was not determined. In other studies with different SMC types, it has been suggested that cAMP interacts with intracellular Ca 2+ signaling pathways at multiple sites [17]. These include a decrease in Ca 2+ influx [3,15,16,18], particularly by the activation of large conductance Ca 2+ -activated K + channels (BK Ca ) to hyperpolarize the membrane [19][20][21], an increase in Ca 2+ efflux [22] or uptake by the sarcoplasmic reticulum [23,24], or an inhibition of agonist-induced IP 3 formation [5,25]. All of these effects of cAMP would be expected to decrease the available intracellular Ca 2+ concentration ([Ca 2+ ] i ). However, it is unknown how any of these effects would influence the frequency of Ca 2+ oscillations of SMCs.
To explore the hypothesis that cAMP induces intrapulmonary airway relaxation by down-regulating Ca 2+ oscillations, we examined the effect of isoproterenol (ISO) and other cAMP elevating agents (FSK and 8-bromo-cAMP) on methacholine (MCh)-induced contractility and Ca 2+ signaling of airway SMCs in murine lung slices. The importance of using lung slices, instead of cultured or isolated cells, is that Ca 2+ signaling and airway contractility can be measured simultaneously. We found that cAMP induced relaxation of MCh-contracted airways and reduced the frequency of the Ca 2+ oscillations. In addition, we confirmed that cAMP did not affect Ca 2+ influx or the refilling of depleted Ca 2+ stores. By contrast, we found that an elevation of intracellular IP 3 concentration ([IP 3 ] i ) reversed the effect of FSK by increasing the frequency of the Ca 2+ oscillations to induce contraction. In addition, we also show that cAMP inhibits the Ca 2+ release via the IP 3 receptor. From these data, we suggest that airway relaxation induced by cAMP is achieved by decreasing the Ca 2+ oscillation frequency of airway SMCs by inhibiting IP 3induced Ca 2+ mobilization.
School. The trachea was exposed and cannulated. After the thoracic cavity was opened, the lung lobes were slowly inflated with a warm agarose solution (2%, 37°C, ~1.3 ml). Subsequently ~0.2 ml of air was injected to flush the agarose out of the airway into the distal alveoli. The lungs were cooled at 4°C to gel the agarose. A single lung lobe was removed and sectioned with a vibratome (model EMS-4000; Electron Microscope Sciences, EMS) into slices of ~140 µm thick starting at the lung periphery. Airways were identified by their lining of epithelial cells with beating cilia. Serial sections were collected and transferred to culture solution containing DMEM supplemented with antibiotics and anti-mycotics and NaHCO 3 . Slices were maintained in culture media at 37°C and 10% CO 2 for up to 3 days.

Measurement of airway contraction and relaxation
Lung slices were observed in a custom-made perfusion chamber constructed from 2 cover-glasses sealed with thin strips of silicone grease (Valve lubricant and sealant, Dow Corning, Midland, MI) on a Plexiglas support. The slices were held in place with a small sheet of nylon mesh (CMN-300-B, Small Parts Inc, Miami Lakes, FL) with a small hole to view the selected airway and avoid the mesh from influencing the contraction of the airway. A gravitydriven perfusion system with a multi-tube manifold with a single output (Warner Instruments, Inc.) was used, in conjunction with valves (LFVA, Lee Company, CT) under TTL control, to exchange solutions in the chamber. The chamber volume was about 100 µl. The perfusion rate was 800 µl/min.
The lung slices were observed with an inverted microscope (IX71, Olympus) with a 10× objective and a zoom adapter. Phase-contrast images were recorded using a CCD camera (model LCL-902C, Watec America Corp, USA), a camera interface ("Picolo", Eurosys Inc., Belgium) and image acquisition software ("Video Savant"; IO industries, Inc, London, Ontario, Canada). Digital images were recorded in time-lapse (0.5 Hz) and analyzed with "NIH /Scion Image" software (Scion Corp.). The lumen area of the airway was measured by summing the number of pixels below a selected threshold grey level with respect to time. All experiments were performed at room temperature (RT, 20 ± 2°C).

Measurements of [Ca 2+ ] i
Slices were loaded for 30 minutes at 30°C with 20 µM Oregon Green 488 BAPTA-1 AM in sHBSS containing 0.1% pluronic (Pluronic F-127; Calbiochem) and 100 µM sulfobromophthalein. Lung slices were de-esterified for another 30 minutes at 30°C in sHBSS containing 100 µM sulfobromophthalein before being placed in the perfusion chamber as described above.
Fluorescence imaging was performed using a custom-built video-rate confocal or 2-photon scanning laser microscope. Either a 488 nm beam from a diode laser or an 800 nm beam from a Ti-sapphire laser (Tsunami, Spectra-Physics, Mountain View, CA) pumped with a 5 W 525 nm diode laser (Millennia, Spectra-Physics) was scanned across the specimen with two oscillating mirrors (X-and Y-scan) through an inverted microscope (Nikon DIA-PHOT 200 for confocal microscopy; Olympus IX71FVSF-2 for 2-photon microscopy). For confocal microscopy, the emitted fluorescence (>510 nm) was separated from the excitation light by a dichroic mirror, a long pass filter and a confocal aperture [26]. For 2-photon microscopy, the emitted fluorescence was separated with a dichroic mirror (670uvdclp, Chroma Technology Corp, Rockingham, VT) and a long pass filter (E700SP or E600SP-2P, Chroma Technology Corp) positioned immediately below the objective. Emitted fluorescence was detected with a photomultiplier tube (PMT, R5929, Hamamatsu USA, Bridgewater, NJ). Grayscale images were recorded at 1, 15 or 30 Hz with a frame-grabber (Raven; Bit Flow, Inc). Changes in fluorescence intensity were obtained, frame-by-frame, from selected regions of interest (ROI, ~5 × 5 pixel). A line-scan analysis was made by extracting a row of pixels from each frame and aligning them in a time sequence.
7.0, Microcal Software Inc, MA, USA). Statistical signifi- The contractile response of airways in mouse lung slices to MCh Figure 1 The contractile response of airways in mouse lung slices to MCh. (A) Low magnification phase-contrast images (scale bar = 50 µm) of the same airway before and 5 minutes after stimulation with 1 µM MCh. L, airway lumen; ECs, epithelial cells; Av, alveolar tissue. (B) The relative change in the area of the airway lumen (% of the initial area) in response to MCh from 1 nM to 1µM. During the interval between each application of MCh (15 min) the airway area was not recorded but the slice was continually washed with sHBSS. (C) The concentration-dependent contractile response of airways to MCh. Each point represents 5 to 7 experiments (mean ± S.E.) in different airways from at least 3 mice. The data were fitted with a logistic function.  cance was determined by using paired student's t-test.

Results
Experiments were only performed on airways that were lined with epithelial cells (ECs) showing active ciliary beating and had a lumen that was free of agarose (Fig. 1A).
To assess the contractile activity of the airway SMCs, the changes in the airway area were measured with time-lapse video microscopy and image-analysis. Upon exposure to MCh, at concentrations ≥ 100 nM, the airway quickly contracted. The luminal area was reduced to, and maintained at, a smaller stable size. Although lower concentrations of MCh (10 nM) induced a reduction in the airway lumen, this change was the result of twitching rather than a sustained contraction of the SMCs (Fig. 1B). The MChinduced airway contraction was concentration-dependent over the range of 10 to 500 nM (Fig. 1C). A maximal contraction of about 52 % was achieved when the concentration of MCh was >0.5 µM. The EC 50 was 78 nM.

The relaxing effect of ISO on airways contracted by MCh
The relaxing effect of ISO on airways was established by initially contracting the airways with MCh for 3 minutes and subsequently exposing the airways to ISO in the presence of MCh. In the first set of experiments, the airways were contracted with a range of MCh concentrations (10 nM to 1 µM) and relaxed with a single concentration of ISO (10 µM). The increasing concentrations of MCh induced an increasing airway contraction and under each condition, addition of ISO induced airway relaxation during the first 2 minutes ( Fig. 2A and 2B). This relaxation effect was not sustained even though the airways were continually perfused with MCh and ISO. After 2 minutes, the airways gradually re-contracted until the lung slice was washed with sHBSS at which time the airway fully relaxed. The magnitude of the initial relaxation, measured after 60 to 90 seconds of ISO exposure, decreased from ~68% tõ 34% (% maximal contraction pre-exposed to ISO) as the concentrations of MCh increased from 10 nM to 1 µM ( Fig. 2B and 2C). The relaxation response was significant different between MCh concentrations (n = 4 slices from 3 mice, p < 0.05). These results indicated that the efficacy of ISO at relaxing the airway is reduced by an increasing contractile stimulus.
Because MCh can stimulate the M 2 receptor and this could possibly lead to the inhibition of adenylyl cyclase (AC) to enhance the contractile response, we examined the effects of the M 2 receptor antagonist, methoctramine (0.2 µM) on ISO-induced relaxation. In the presence of the M 2 antagonist, airway contraction induced by MCh (200 nM) was decreased by about ~5%. However, the transient relaxation induced by ISO (10 µM) was unchanged in the presence of methoctramine. A similar transient relaxation was also observed in response to ISO (10 µM), when the airway was pre-contracted with another agonist, serotonin (5-HT, 1 µM).
In a second set of experiments, the airways were contracted with a single concentration of MCh and relaxed with different ISO concentrations. The previous data indicated that it was necessary to use a low MCh concentration in order that the relaxation responses could be differentiated. We determined that 200 nM MCh satisfied these requirements. In airways contracted with MCh (200 nM), ISO induced relaxation with a typical pattern of an initial relaxation followed by a slow re-contraction ( Fig. 3D solid line). The magnitude of the initial relaxation was concentration-dependent from 1 nM to 1 µM ISO, ranging from 11 ± 1% (n = 5 slices from 3 mice) to 53 ± 2% (n = 6 slices from 4 mice) ( Fig. 3E solid line). No further relaxation was induced by ISO at concentrations higher than 1µM (Fig. 3E).

Ca 2+ events underlying airway contraction and relaxation
In response to 200 nM MCh, the airway SMCs demonstrated an initial increase of intracellular Ca 2+ which was followed by the onset of Ca 2+ oscillations (Figs. 3B and 3C). These Ca 2+ oscillations were superimposed on a slightly elevated baseline. The frequency of the Ca 2+ oscillations increased to a stable rate within 1 minute. Upon exposure to 1 µM ISO, the frequency of the Ca 2+ oscillations slowed, although the baseline of the Ca 2+ signal did not substantially change. In the continued exposure to ISO with MCh and after about 150 seconds, the frequency of Ca 2+ oscillations began to increase and approached a new steady rate; this rate was less than the initial rate induced by MCh alone.
The frequency of the Ca 2+ oscillations was determined from the period of each oscillation (time interval between 2 adjacent peaks of the Ca 2+ signal). As a result a correlation between the dynamic changes in the frequency of the Ca 2+ oscillations with the changes in the cross-sectional area of the airway could be made (Fig. 3D). In response to MCh, the airway contracted to a smaller stable size and this correlated with the initiation and increase in the frequency of the Ca 2+ oscillations to a stable rate. Similarly, the ISO-induced decrease in the frequency of the Ca 2+ oscillations correlated with the ISO-induced initial relaxation of the airway. The subsequent re-contraction of the airway correlated with the late increase in the frequency of the Ca 2+ oscillations.
The effect of ISO on the frequency of the Ca 2+ oscillations was expressed as the ratio of the frequency of the Ca 2+ oscillations after ISO exposure (60 to 90 seconds) to the frequency of the Ca 2+ oscillations before ISO exposure. In the range from 1 nM to 10 µM, ISO induced a concentration-dependent decrease in the frequency ratio from 0.93 ± 0.03 (n = 5 slices from 3 mice) to 0.56 ± 0.03 (n = 6 slices from 4 mice). The decline in the frequency ratio was inversely proportional to the change in airway size (Fig.  3E).

Relaxation of airways by repeated exposure to ISO
Although we found that ISO induced a concentrationdependant relaxation of pre-contracted airway, this relaxation response did not persist in the presence of ISO. A possible explanation for this response is receptor accommodation or desensitization to the continual presence of agonists. To address this possibility, we examined the relaxation response of the same contracted airway (with MCh, 200 nM) to multiple exposures of ISO (1 µM) (Fig.  4). The interval between each ISO exposure was increased from 10 minutes to 2 hours and 18 hours. In comparison with the initial relaxation response to ISO, the subsequent relaxation of the airway decreased to 50 ± 9 % (n = 5 slices from 4 mice, p < 0.01) in response to ISO exposure after 10 minutes (Fig 4A and 4B). However, the efficacy of ISO was substantially improved when the interval between exposures was lengthened to 2 hours. After 18 hours, the ISO induced-relaxation was found to have recovered to 97 ± 3 % (n = 2 slices from 2 mice, P = 0.36 compared to initial effect) ( Fig. 4A and 4B).

Relaxing effect of FSK and 8-bromo-cAMP on contracted airways
To investigate the airway relaxation response to the elevation of [cAMP] i , we examined the effects of FSK, an activator of AC, and 8-bromo-cAMP, an active membrane permeable analogue of cAMP. Both FSK (10 µM) and 8bromo-cAMP (500 µM) induced an initial fast relaxation followed by a slow progressive relaxation of MCh contracted airways ( Fig. 5A and 5D solid line). The increase in the lumen area induced by 500 µM 8-bromo-cAMP was 49 ± 9 % (n = 3 slices from 2 mice). The effect of FSK was concentration-dependent (Fig. 5C); increasing concentrations of FSK (10 nM to 10 µM) induced an increasing relaxation (measured after 5 minutes exposure) from 4 ± 1 % to 75 ± 2 % (n = 4 slices from 3 mice).
Recordings of intracellular Ca 2+ in airway SMCs revealed that both FSK and 8-bromo-cAMP reduced the frequency of the Ca 2+ oscillations initiated by MCh ( Fig. 5B and 5E). In contrast to the effects of ISO, the frequency of the Ca 2+ oscillations declined throughout the experiment. In range from 10 nM to 10 µM, FSK induced a concentrationdependent decrease in the frequency of the Ca 2+ oscillations (Fig. 5C). The frequency of the Ca 2+ oscillations after 5 minutes of FSK exposure, expressed as the ratio to the Ca 2+ oscillation frequency induced by MCh exposure, decreased from 0.95 ± 0.03 to 0.42 ± 0.04 (n = 4 slices from 3 mice) as the FSK concentration increased. The changes in frequency of the Ca 2+ oscillations induced by FSK or 8-bromo-cAMP correlated with a progressive relaxation of the airway (Fig. 5A and 5D).
The ability of ISO to relax an airway is influenced by repeti-tive stimulation  The effect of FSK and 8-bromo-cAMP on contraction and Ca 2+ oscillations of airway SMCs

A decrease in Ca 2+ oscillation frequency correlates with airway relaxation
During all of the relaxation responses to ISO, FSK or 8bromo-cAMP, the frequency of the Ca 2+ oscillations initiated by MCh was reduced (Fig. 3D, Fig. 5A and 5D). A summary plot of these data (from Fig. 3E, Fig. 5C and 5D) is presented in figure 6. This analysis suggests that, irrespective of the agonist, the frequency of the Ca 2+ oscillations determines the extent of airway relaxation: the lower the frequency of the Ca 2+ oscillations in airway SMCs, the greater the relaxation of the airway (Fig. 6).

cAMP does not affect caffeine-induced Ca 2+ release
To explore the mechanisms by which cAMP reduced the frequency of the Ca 2+ oscillations, we investigated the effect of cAMP on caffeine-induced Ca 2+ release and refilling of intracellular Ca 2+ stores (Fig. 7). In view of the transient nature of ISO-induced relaxation, FSK was used to elevate cAMP because of its direct and stable action on AC. Upon exposure to 20 mM caffeine, the airways responded with a transient contraction (Fig 7A). This treatment also induced in the SMCs a transient increase in [Ca 2+ ] i (peak fluorescence intensity ratio, F/F 0 = ~3.5) followed by a sustained plateau of [Ca 2+ ] i (fluorescence ratio = ~1.8, Fig.  7D). The incubation of the same airway with FSK (10 µM for 5 min) had no significant effect on caffeine-induced Relationship between the reduction in Ca 2+ oscillation fre-quency in SMCs and airway relaxation Figure 6 Relationship between the reduction in Ca 2+ oscillation frequency in SMCs and airway relaxation. The relaxation state (percentage of the MCh-induced contraction) and corresponding Ca 2+ oscillation frequency (ratio to the frequency of MCh-induced Ca 2+ oscillations) upon exposure to ISO, FSK and 8-bromo-cAMP (at various concentrations) were plotted and fitted with a logistic function. The data indicates that a greater airway relaxation correlated with a larger reduction in the Ca 2+ oscillation frequency.   (Fig. 7E). The ratio of contraction induced by the second exposure to caffeine as compared to the first exposure was 0.96 ± 0.02 in the absence of FSK and 0.88 ± 0.04 in the presence of FSK (Fig. 7C). The ratios of the peak fluorescence intensities were 0.96± 0.06 in the absence of FSK and 1.0 ± 0.06 in the presence of FSK (n = 4 slices from 3 mice, Fig. 7F). The similarity of the results suggests that cAMP does not affect caffeine-induced Ca 2+ release (i.e. via the ryanodine receptor) from the SR or the subsequent refilling of the depleted internal Ca 2+ store.

Supplementation of intracellular Ca 2+ does not block the relaxing effect of cAMP
Because a reduction of [Ca 2+ ] i , brought about by decreasing Ca 2+ influx and/or increasing Ca 2+ efflux or reuptake, has been proposed as a mechanism by which cAMP relaxed airway SMCs, we examined the effect of elevating [Ca 2+ ] i during the response of SMCs to cAMP. The [Ca 2+ ] i was increased with 3 different agents: 1) NP-EGTA, a caged-Ca 2+ compound, which upon UV photolysis releases Ca 2+ , 2) ionomycin, a Ca 2+ selective ionophore, and 3) iberiotoxin (IbTX), a specific inhibitor of BK Ca channels, which enhances Ca 2+ influx by blocking K + efflux to induce hyper-polarization.
In airway SMCs pre-treated with MCh (200 nM) and FSK (10 µM), flash photolysis of caged-Ca 2+ within the SMC increased the fluorescence intensity of intracellular Ca 2+ signaling (Fig. 8B). However this served only to elevate the baseline, but not the frequency, of the Ca 2+ oscillations ( Fig. 8C and 8D, 6 airways from 4 mice). No change in the airway luminal area or an increase in contraction of the illuminated SMCs was observed.
Upon exposure to ionomycin (5 µM), the relaxation induced by FSK (10 µM) in MCh (200 nM) contracted airway SMCs was initially enhanced, reaching a maximal relaxation after 2 min (Fig. 9A). This additional relaxation correlated with the cessation of the Ca 2+ oscillations ( Fig.  9B and 9C). After continued ionomycin exposure, the airway began to re-contract and this correlated with an steady increase in the [Ca 2+ ] i (Fig 9A and 9B) (5 airways from 4 mice).
A similar change in the contraction and [Ca 2+ ] i of airway SMCs was observed when the lung slices treated with MCh (200 nM) and FSK (10 µM) were subsequently exposed to IbTX (50 nM). The airway responded with a quick and full relaxation within 3 mins (Fig. 10A). Again, the accompanying change in the [Ca 2+ ] i revealed a rapid slowing of the frequency of the Ca 2+ oscillations to a full stop. In contrast to the effect of ionomycin, the [Ca 2+ ] i of SMCs remained near baseline after the Ca 2+ oscillations ceased (Fig. 10B, 6 airways from 5 mice).  A similar set of experiments was then performed in the presence of MCh and FSK. The inhibition of the frequency of the Ca 2+ oscillations and the accompanying relaxation of SMCs were observed when the slices were exposed to MCh (200 nM) and FSK (10 µM). But, in response to UV illumination (0.5s, OD 0.5) and photolysis of caged IP 3 , the frequency of the Ca 2+ oscillations increased significantly for a short period (Fig. 11B, C, D). A second flash again increased the oscillation frequency, although to a lesser extent. Simultaneous recording of changes in the frequency of Ca 2+ oscillations and contraction of airway SMCs revealed that the airway re-contracted after each UV flash (Fig. 11C). These results suggest that FSK exerts its relaxant effect by inhibiting the IP 3 induced Ca 2+ release from internal Ca 2+ stores.

Effect of iberiotoxin on FSK-induced airway relaxation
If the IP 3 receptor is inhibited by cAMP, it should be possible to block Ca 2+ release induced by IP 3 in cAMP treated SMCs. In control experiments, a lower level of UV illumination (reduced to 1/3, by using a neutral density filter with OD 1.0) induced a single spike of elevated [Ca 2+ ] i in untreated airway SMCs (Fig. 12). A second flash gave a similar response. However, after the SMCs were treated with FSK (10 µM) for 5 minutes, the increase in [Ca 2+ ] i was virtually abolished in response to the UV flash ( Fig.  12B and 12C).

CICR through the ryanodine receptor is not required for Ca 2+ oscillations
Because the ryanodine receptor (RyR) is involved in Ca 2+induced Ca 2+ release (CICR) in a variety of cell types, we examined the effect of CICR via RyR in airway SMCs. We Effect of FSK on the intracellular Ca 2+ signaling induced by IP 3 Figure 12 Effect of FSK on the intracellular Ca 2+ signaling induced by IP 3 . The Ca 2+ signaling of SMCs induced by repetitive flash-photolysis (0.5 s, OD 1.0, arrows) of caged-IP 3 in lung slices incubated (A) without or (B) with FSK for 5 min. Exposure to FSK almost completely abolished the Ca 2+ response induced by flash photolysis of IP 3 . (C) Summary of the changes in Ca 2+ signaling (peak fluorescence intensity) induced by IP 3 in the slices incubated without or with FSK. The Ca 2+ signal after exposure to the sHBSS without or with FSK was expressed as a ratio to the Ca 2+ response before exposure. Data represents the mean ± S.E. from at least 5 different slices from 4 mice. *, p < 0.05. found that exposure to ryanodine (50 µM, sufficient to inhibit the RyR) induced only a small and slow decrease in the MCh-induced airway contraction (~5%, after 5 minutes, Fig. 13A) and frequency of Ca 2+ oscillation in the SMCs (~4 min -1 , after 5 minutes, Fig. 13B). The amplitude of the Ca 2+ oscillations was unaffected. These result suggest that the increases in the [Ca 2+ ] i associated with the Ca 2+ oscillations are not mediated by RyR or are sufficient to invoke CICR via the RyR.
To examine the Ca 2+ conditions required for the induction of CICR, we released Ca 2+ by flash photolysis of caged Ca 2+ . When a restricted area within a single SMC was exposed to UV light (Fig 13C), only a brief localized increase in [Ca 2+ ] i was detected (Fig. 13D). The increase in Ca 2+ , as determined by the fluorescence change, was similar to that induced by a Ca 2+ oscillation, but this Ca 2+ change did not spread to the rest of the cell. By contrast, when the same airway SMCs was treated with sHBSS containing 50 mM K + for ~3 minutes, a second identical UV flash induced a larger Ca 2+ release in the zone of illumination and induced a Ca 2+ wave that propagated throughout the cell (Fig 13D, 5 airways from 3 mice). Treatment with high K + was used because it has previously been reported that high K + initiates membrane-depolarization that results in excessive Ca 2+ influx, overfilling of internal Ca 2+ store and sensitization of the RyR [14]. These results indicate that CICR through RyR only occurs in airway SMCs under certain circumstances or "stressed conditions" and implies that RyRs do not play a significant role in the normal maintenance of agonist-induced Ca 2+ oscillations.

Discussion
In this study, thin lung slices were used to explore the changes in intracellular Ca 2+ that occur in airway SMCs during relaxation induced by ISO and other cAMP elevating compounds. The airway SMCs were initially contracted with MCh, an analogue of ACh, which is more slowly metabolized and widely used in clinical lung function tests. We found that MCh induced both a concentration-dependent contraction of the airways and increase in the frequency of Ca 2+ oscillations within airway SMCs. These results are similar to our previous findings obtained with ACh in lung slices [12,14]. Upon exposure to ISO, the MCh-contracted airways initially displayed a transient relaxation but subsequently re-contracted. The ability of ISO to relax the airway was dependent both on the ISO concentration and the MCh concentration used to contract the airway. If the airways were contracted with high concentration of MCh (1 µM), ISO, even at high concentrations (i.e.10 µM) was ineffective at relaxing the airway. Only at low MCh concentrations was ISO able to induce substantial relaxation. A similar relaxation and subsequent re-contraction in response to ISO (10 nM ~10 µM) was also observed in MCh-contracted (1 µM) rat lung slices (data not shown). These results suggest that the relaxing action of ISO is relatively weak in comparison to the contractile action of MCh and that the airway becomes de-sensitized to ISO-induced relaxation to allow re-contraction.
In addition to being activated by β-adrenoceptor associated G-stimulatory proteins (G s ), AC can also be inhibited by receptor-coupled G-inhibitory proteins (G i ). Therefore, a potential mechanism for the down-regulation of the ISO response is the inhibition of AC by activation of M 2 muscarinic receptors and their associated G i by MCh [28,29]. However, in the presence of the M 2 receptor antagonist, methoctramine, MCh only induced a slightly smaller airway contraction (about 95%), but more importantly, ISO induced a similar transient relaxation. This transient response to a high concentration of ISO also persisted across a wide range of MCh concentrations, which would be expected to differentially activate the M 2 receptors. Moreover, a similar transient response to ISO was also observed in the airways contracted by 5-HT, an agonist with no effect on the AC. In other studies [30,31], the pretreatment of cultured SMCs or bronchial smooth muscle strips with ISO (30 min to 24 hours) reduced subsequent ISO-mediated cAMP generation and relaxation. Similarly, we found that repetitive exposure of the airway to ISO (1 µM, 10 min to 2 hours) significantly decreased the relaxation response (by 30 to 50%). These results indicate that the M 2 receptor has a minimal effect on AC activity at midrange MCh concentrations and that the down-regulation of the ISO response is mediated by rapid de-sensitization and slow re-sensitization at the β-adrenoceptor. Desensitization of the β-adrenoceptor can be mediated by phosphorylation of the receptor by cAMP-independent βadrenoceptor kinase or other G protein-coupled receptor kinases or by receptor internalization -a process that takes a longer time to reverse (hours) [2,32].
Because β-adrenoceptor agonists primarily relax SMCs by increasing [cAMP] i , an implication of receptor desensitization is that the [cAMP] i within the SMCs initially increases but subsequently declines. The idea that increases in [cAMP] i correlate with relaxation was addressed by circumventing the β-adrenergic receptor by either activating adenylyl cyclase with FSK or by increasing the [cAMP] i directly with 8-bromo-cAMP. A major characteristic of FSK or 8-bromo-cAMP-induced relaxation was that the relaxation was sustained and that re-contraction never occurred. This response is consistent with the idea that the [cAMP] i must be maintained or increased if relaxation is to be sustained and that some form of desensitization of β-adrenoceptor occurs allowing re-contraction.
The correlation of increases in [cAMP] i with airway relaxation has been reported in mice [33], rats [23,34], guinea pigs [35], rabbits [36,37], dogs [3], cows [4,24,38], and humans [24,37]. However, these studies were all conducted on isolated airway muscle strips or tracheal rings by measuring changes in isometric tension. Consequently, to explore the cellular mechanisms by which ISO and other agents acting via cAMP to relax SMCs, we examined their effects on intracellular Ca 2+ . As previously described, MCh, as well as ACh, initiates Ca 2+ oscillations in SMCs and the frequency of these Ca 2+ oscillations determines the extent of the SMC contraction [9,11,12,14]. During the initial response to ISO, the frequency of the MCh-induced Ca 2+ oscillations was quickly reduced and this was accompanied by fast relaxation of the airway. However, after a short time, the Ca 2+ oscillation frequency began to slowly increase and this was accompanied by the slow re-contraction of the airway. A similar slowing of the frequency of the Ca 2+ oscillations and airway relaxation was induced by FSK and 8-bromo-cAMP. But, in contrast to ISO, the frequency of the Ca 2+ oscillations continued to slow to lower frequencies; a result that correlates with the greater airway relaxation.
Although the relationship to contraction could not be measured, Salbutamol, a β 2 -adrenoceptor agonist, has also been found to slow and reduce the baseline of AChinduced Ca 2+ oscillation in dissociated porcine tracheal SMCs [16]. In our studies, we found that the baseline of the Ca 2+ oscillations was relatively stable. Consequently, we hypothesized, especially in view of the fact that the peak [Ca 2+ ] i of the Ca 2+ oscillations is substantially high, that increases in [cAMP] i induce airway relaxation by reducing the frequency of Ca 2+ oscillations in SMCs rather than by lowering the baseline of global [Ca 2+ ] i .
We [12][13][14] and others [10,39]  Another mechanism inhibiting SMC Ca 2+ influx appears to involve membrane hyper-polarization resulting from cAMP-activation of BK Ca channels to down-regulate voltage-dependent Ca 2+ influx [19][20][21]. Specific blockers of the BK Ca channels such as IbTX or charybdotoxin have been shown to antagonize cAMP-induced relaxation in isolated trachea or bronchi from guinea-pigs [21], pig [40], horse [41]. However, in keeping with the effects of ionomycin, we found that in lung slices treated with MCh and FSK, IbTX further inhibited the Ca 2+ oscillations to enhance airway relaxation. This effect suggests IbTX (and ionomycin) may actually inhibit Ca 2+ oscillation by enhancing Ca 2+ influx. The failure to observe a significant elevation of baseline Ca 2+ suggests either that the influx of Ca 2+ is sufficiently low and quickly sequestered into internal stores or that an alternative, non-specific action of IbTX must be considered. The failure of IbTX to induce recontraction is consistent with the fact that voltagedependent Ca 2+ influx does not appear to play a prominent role in maintaining the agonist-induced Ca 2+ oscillations in airway SMCs and that airway contraction and Ca 2+ signaling induced by membrane depolarization (with KCl) were completely different to that induced by agonists [42]. Furthermore, there are conflicting reports regarding the contribution of BK Ca channels based on differences amongst species and regions of the airway as well as experimental conditions [43,44]. Consequently, it appears unlikely that the BK Ca channel is a major mechanism mediating the effects of cAMP in mouse airway SMCs. Taken together, these data led us to believe that decreased [Ca 2+ ] i is not the primary cause of the modulation of the frequency of Ca 2+ oscillations.
Therefore, we examined the third idea that cAMP altered the Ca 2+ release from SR. Although the RyR has an important role in CICR in many other cell types [45], Ca 2+ oscillations and airway contraction induced by MCh were minimally affected in the short term by the presence of ryanodine. This suggests that the Ca 2+ oscillations do not require RyRs and therefore it is unlikely that any effects of cAMP on RyR will be significant. The slight decline in the Ca 2+ oscillation frequency may be explained by an increased leak of Ca 2+ through RyR locked in the open state by ryanodine. However, this process must be occurring independently of the Ca 2+ oscillations otherwise the Ca 2+ oscillations would quickly stop as a result of empty Ca 2+ stores. These results are also consistent with the fact that the release of caged-Ca 2+ only initiated CICR when the cell was previously depolarized with KCl to overload the internal Ca 2+ stores and sensitize the RyR.
The IP 3 R is another important Ca 2+ release channel of the SR and we found that the slow frequency of Ca 2+ oscillations induced by FSK could be increased when the [IP 3 ] i was increased by flash photolysis. This response was accompanied by the re-contraction of airway SMCs. The idea that the sensitivity of the IP 3 R to IP 3 is down-regulated by FSK is further supported by the finding that photolysis of caged IP 3 induced a propagating Ca 2+ wave in the absence, but not in the presence of FSK. A similar effect of cAMP on Ca 2+ release has also been reported in other types of SMCs or non-excitable cells [46][47][48][49]. The underlying mechanism may involve cAMP-dependent PKA to phosphorylate the IP 3 R. However, the outcome of phosphorylation differs with the preparation and either an enhancement or inhibition of Ca 2+ release can result [46][47][48][49]. Alternatively, cAMP may influence the IP 3 R via EPACs.
Although the molecular details mediating the action of cAMP will require further investigation, we propose that the cAMP has an inhibitory effect on the IP 3 [51,52]. In view of this change in sensitivity, we hypothesize that the inhibition of IP 3 R by cAMP results in a similar leftward and downward shift of the bell-shaped curve. This idea is consistent with our observation that in the presence of cAMP, an increase in Ca 2+ influx, induced by IbTX or ionomycin, leads to the closure of the IP 3 R to mediate the cessation of Ca 2+ oscillations and relaxation of the airway SMCs.
An additional mechanism that would augment the inhibitory action of cAMP on the IP 3 R is a reduction in the production of IP 3 . This idea is based on the evidence that G protein-activated PLCβ activity can be inhibited by PKAinduced phosphorylation [4,5,53,54]. This requires further investigation but is beyond the scope of this study.
An alternative mechanism for cAMP-induced SMC relaxation is the desensitization of contractile apparatus to Ca 2+ [8,24,40,55]. Phosphorylation of MLCK can lead to decreased affinity of MLCK for Ca 2+ /calmodulin while phosphorylation of Rho-kinase or protein kinase C can increase the activity of MLCP. We have also found that cAMP can down-regulate the Ca 2+ -sensitivity of agonistinduced contraction in lung slices [56]. However, it is not yet clear if such changes occur in parallel with the changes in Ca 2+ oscillation frequency.
In summary, we conclude that the frequency of Ca 2+ oscillations controls the contractile state of airway SMCs and that compounds that elevate cAMP slow the frequency of Ca 2+ oscillations within SMCs to proportionally induce airway relaxation. Furthermore, cAMP appears to regulate the Ca 2+ oscillation frequency by inhibiting Ca 2+ release through IP 3 R rather than influencing Ca 2+ influx or the refilling of internal stores. In addition, we find that, under normal conditions, the RyR is not essential for Ca 2+ oscillations of airway SMCs.