Ryanodine receptor sensitization results in abnormal calcium signaling in airway smooth muscle cells.

AJRCMB Articles in Press. Published on 15-April-2015 as 10.1165/rcmb.2014-0386OC

Page 1 of 36

Ryanodine receptor sensitization results in abnormal calcium signaling in airway smooth muscle cells Huguette Croisier1 , Xiahui Tan2 , Jun Chen3 , James Sneyd4 , Michael J. Sanderson3 , and Bindi S. Brook1∗

1

School of Mathematical Sciences, University of Nottingham, Nottingham, United Kingdom

2

Lung Inflammation and Infection Laboratory, QIMR Berghofer Medical Research Institute, Her-

ston, Queensland, Australia 3

Department of Microbiology and Physiological Systems, University of Massachusetts Medical

School, Worcester, MA, USA 4

Department of Mathematics, University of Auckland, Auckland, New Zealand

∗

Corresponding author; [email protected], +44(0)115 9513127

Authors contributions: Conception and design: HC, MJS, JS, BSB; Analysis and Interpretation: HC, XT, BSB, JS; Drafting the manuscript for important intellectual content: HC, MJS, JS, BSB; Performed the experiments: XT, JC, MJS

Running title: RyR sensitization alters Ca2+ signaling in ASMCs

1 Copyright © 2015 by the American Thoracic Society


Abstract Intracellular Ca2+ dynamics of airway smooth muscle cells (ASMCs) are believed to play a major role in airway hyper-responsiveness and remodeling in asthma. Prior studies have underscored a prominent role for IP3 receptors (IP3 Rs) in normal agonist-induced Ca2+ oscillations, while ryanodine receptors (RyRs) appear to remain closed during such Ca2+ oscillations, which mediate ASMC contraction. Nevertheless, RyRs have been hypothesized to play a role in hyperresponsive Ca2+ signaling. This could be explained by RyRs being “sensitized” to open more frequently by certain compounds. We investigate the implications of RyR sensitization on Ca2+ dynamics in ASMC using a combination of mathematical modeling and experiments with mouse precision cut lung slices (PCLS). Caffeine is used to increase the sensitivity of RyRs to cytosolic Ca2+ ([Ca2+]i) and sarcoplasmic reticulum Ca2+ ([Ca2+ ]SR ). In ASMC, large caffeine concentrations (>10mM) induce a sustained elevation of [Ca2+ ]i. Our mathematical model accounts for this by the activation of store-operated Ca2+ entry (SOCE) that results from a large increase in the RyR sensitivity to [Ca2+ ]SR and the associated Ca2+ release which leads to a reduction of [Ca2+ ]SR . Importantly, our model also predicts 1) that moderate RyR sensitization induces slow Ca2+ oscillations, a result experimentally confirmed with low concentrations of caffeine; and 2) that high RyR sensitization suppresses fast, agonist-induced Ca2+ oscillations by inducing substantial SOCE and elevated [Ca2+ ]i. These results suggest that RyR sensitization could play a role in ASMC proliferation (by inducing slow Ca2+ oscillations) and in airway hyperresponsiveness (by inducing greater mean [Ca2+]i for similar levels of contractile agonist).

Keywords:

precision cut lung slice (PCLS), mathematical modeling, asthma, hypersensitivity,

Ca2+ oscillations


Page 2 of 36


Page 3 of 36

1

Introduction

Intracellular Ca2+ dynamics control a wide variety of cellular processes. In particular, airway smooth muscle cell (ASMC) contraction is mediated by oscillations in [Ca2+ ]i , which correspond to the cyclic release and re-uptake of Ca2+ from the sarcoplasmic reticulum (SR). The release of Ca2+ from the SR occurs via two types of channels, the IP3 receptor (IP3 R) and the ryanodine receptor (RyR), while Ca2+ reuptake into the SR is performed by Ca2+ ATP-ases (SERCAs). Ca2+ may also enter the cell via receptor-operated, store-operated, and voltage-operated Ca2+ channels (ROCC, SOCC, and VOCC respectively) and be extruded from the cell across the plasma membrane by Ca2+ ATP-ases (PMCAs). In ASMC, GPCR stimulation (by agonists) and membrane depolarisation (by KCl) induce 2 types of Ca2+ oscillations [1, 2]. KCl-induced Ca2+ oscillations have a low frequency (∼3/min) as compared with faster agonist-induced Ca2+ oscillations (∼10/min in human and ∼30/min in mouse at room temperature; these are even faster at 37deg C). In mouse ASMCs, KCl-induced Ca2+ oscillations induce twitching of the ASMCs and substantially less airway contraction as compared to agonist-induced contraction, but the relative magnitudes of KCl- versus agonist-induced contraction is species-dependent [3, 2]. We have previously [4] proposed a mechanism for how IP3 R and RyR interact to give these two oscillatory behaviors. The pathways are summarised in Figure 1. In brief: • Agonist stimulation, and subsequent production of IP3 , leads to release of Ca2+ through the IP3 R, and a cycle of Ca2+ release and reuptake to and from the SR, via the mechanism of Ca2+ -induced Ca2+ release (CICR). This causes sufficient depletion of the SR to inactivate the RyR, which thus play no significant role in on-going agonist-induced Ca2+ oscillations (Fig. E1A). • KCl stimulation causes over-filling of the SR, which activates the RyR and leads to Ca2+ release from the SR, again in a process of CICR (Fig. E1B). • The IP3 R and RyR interact via their joint effects on Ca2+ concentration in the cytosol and the SR. The fact that RyR get desensitized when the SR depletes plays a crucial role in determining the relative contributions of IP3 R and RyR to the control of Ca2+ oscillations. From the above, increasing the Ca2+ sensitivity of RyR (which we refer to in this work as RyR “sensitization”) could have important implications for the participation of RyRs in Ca2+ dynamics in hyper-responsive conditions. In addition to agonist-induced calcium signalling being intrinsically altered in asthmatic airway (e.g., [5]), pro-inflammatory factors present in asthma enhance [Ca2+ ]i 3 Copyright © 2015 by the American Thoracic Society


responses of (non-asthmatic) ASMCs to agonist stimulation and/or to store depletion [6, 7, 8, 9, 10, 11, 12]. There is evidence that the RyR could play a role in these augmented Ca2+ responses [13, 14]. This role for RyR sensitivity can be explored experimentally with compounds such as caffeine that increase the open probability of RyR to [Ca2+ ]SR and [Ca2+ ]i [15, 16]. Therefore, in this work, we integrate both IP3 R and RyR dynamics into an extension of our previous mathematical model of Ca2+ dynamics in ASMC [17] and experimentally test the predictions of this model. The consequences of RyR sensitization on ASMC Ca2+ dynamics is evaluated in the absence and presence of agonist stimulation.

2 2.1

Material and Methods Experimental methods

Precision cut lung slices (PCLS) were prepared from female BALB/c mice (7-12 weeks old) as described in Supplemental Material (section E1.1). All experiments were conducted at body temperature (37degC) in a custom-made, temperature-controlled microscope chamber as described in [3]. Measurement of Ca2+ oscillations: PCLS were incubated in sHBSS containing 20 µM Oregon Green 488 BAPTA-1-AM (Invitrogen), a Ca2+-sensitive dye, 0.1% Pluronic F-127 (Invitrogen) and 200 µM sulfobromophthalein in the dark at 30degC for 1 hour. Subsequently, the PCLS were incubated in 200 µM sulfobromophthalein for 30 min. Slices were mounted on a cover-glass and held down with 200 µm nylon mesh. A smaller cover-glass was placed on top of the mesh, sealed at the sides with silicone grease to facilitate solution exchange. Mounted lung slices were continuously perfused with sHBSS or sHBSS containing the required compounds. ASMC Ca2+ signals were examined with a custom-built 2-photon scanning laser microscope with a x40 oil immersion objective (NA 1.35) and images recorded at 15 or 30 images s−1 . Changes in fluorescence intensity (which represent changes in [Ca2+ ]i) were analysed in an ASMC of interest by averaging the grey value of up to a 10 x 10 pixel region using custom written software. Relative fluorescence intensity was expressed as a ratio of the fluorescence intensity (Ft /F0 ) i.e. fluorescence intensity at a particular time (Ft ) normalized with respect to the initial fluorescence intensity (F0 ).


Page 4 of 36


Page 5 of 36

2.2

Statistical analysis

√ Results on Ca2+ oscillation frequency are shown as mean ± standard error of the mean (SEM=SD/ n, where SD is the standard deviation, and n the number of observations). Student’s t test was used to compute the 95% confidence interval of the mean (mean ±tn−1 0.975SEM), and to evaluate the significance of the difference between means at different caffeine concentrations.

2.3

Mathematical model

The model is an extension of previous work [17]. In brief, the dynamics of c =[Ca2+ ]i and cs =[Ca2+ ]SR are governed by: dc = Jin − JP M CA + Jrel − JSERCA , dt dcs = γ(J SERCA − Jrel ), dt

(1)

where Jin models Ca2+ influx (principally through store-operated channels), JP M CA and JSERCA are the Ca2+ fluxes through the PMCA and SERCA pumps respectively, and Jrel models the flux out of the SR through IP3 R and RyR. Full details of each of these fluxes are given in the Supplement (section E1.2). For our purposes, the most important term is Jrel , which contains a model of the RyR flux. In this flux, the open probability of the RyR is assumed to be given by PRyR (c, cs ) =

c3 K3

cs4 4

.

(2)

3 4 cR + c KsR + c s

Thus, the RyR sensitivity to cytosolic and luminal Ca2+ is governed by the two parameters KcR and KsR , respectively.

3

Results

3.1 3.1.1

Effect of RyR sensitization on resting Ca2+ Model predictions

Sensitization of the RyR is obtained in the model by decreasing the cytosolic Ca2+ activation threshold, KcR , or the luminal Ca2+ activation threshold, KsR from their resting values (Eq. (2) and Table E1). We define the indices, rc and rs , to reflect the degree of cytosolic and luminal sensitization respectively, as follows: rc = (KcR0 − KcR )/Ke and rs = (KsR0 − KsR) /Ks , where KcR0 and KsR0



are the resting values of KcR and KsR in Table E1. Thus, increasing values of rc and rs represent increasing cytosolic and luminal sensitization of the RyR to Ca2+, respectively. Figs. 2A and B show the effect of increasing respectively rc and rs on [Ca2+ ]i dynamics. Both moderate cytosolic and luminal sensitization of the RyR induce low-frequency RyR-mediated Ca2+ oscillations (panels (i) and (ii)), with a frequency that increases with increased sensitization. We note that the induction of Ca2+ oscillations by an increase in rs can be understood as the mirror mechanism by which KCl induces Ca2+ oscillations : the increase in [Ca2+ ]SR induced by membrane depolarization, which activates the RyR (Fig. E1B), is effectively replaced by a decrease in the [Ca2+]SR half-activation threshold KsR (i.e., overcoming the threshold is equivalent to lowering it sufficiently). At higher RyR sensitization (Fig. 2A,B(iii)), the Ca2+ oscillations disappear and are replaced by a steady [Ca2+ ]i , with one important difference: the steady [Ca2+ ]i induced by high cytosolic RyR sensitization remains close to the resting [Ca2+ ]i , whereas the steady [Ca2+ ]i induced by high luminal sensitization is substantially elevated. This is worth noting for comparison with experimental observation ([1] and Fig. 2C(iii); see below). The model indicates that this difference in steady-state is due to the different levels of SOCE activation. Indeed, as we have previously shown [17], a sustained increase in Ca2+ influx (i.e., a Ca2+ flux from the extracellular medium into the cytosol) is required for a sustained increase in [Ca2+ ]i. An increase in RyR luminal sensitivity (rs), but not cytosolic sensitivity (rc ), induces a sustained Ca2+ influx because it allows RyR to remain open at lower [Ca2+]SR and thus SOCE is substantially activated. Fig. 2E illustrates this mechanism: with increased RyR luminal sensitivity, the RyRs open more frequently (magenta curve) because the [Ca2+]SR (Fig. 2D) reaches the lower threshold for RyR activation sooner. Simultaneously, because the [Ca2+]SR remains lower, a larger fraction of SOCC remain open (Fig. 2E: blue curve). The results of Figure 2A, B are summarised and extended in the Supplement (Figure E3). 3.1.2

Experimental validation

Caffeine has been shown by other groups to increase RyR sensitivity to both cytosolic and luminal Ca2+ in artificial lipid bilayers [15, 16]. In addition, in ASMC, high caffeine concentrations (20mM) induce a sustained, elevated [Ca2+ ]i [1], which is a signature for luminal RyR sensitization according to our model (Fig. 2B(iii)). Consequently, to validate the predictions of our model that RyR sensitization influences the Ca2+ dynamics of ASMCs, we used varying concentrations of caffeine to sensitize the RyR of ASMCs in mouse lung slices (Fig. 2C).


Page 6 of 36


Page 7 of 36

Stimulation of ASMCs with 1 - 4mM caffeine (at 1 mM intervals) induced low frequency [Ca2+ ]i oscillations (Figs. 2C, (i )and (ii)). The experimental dose-response curve for caffeine (for a total of 4 different caffeine concentrations) is given in the Supplementary Material, Figure E4. It can be seen that the oscillation frequency increases as a function of caffeine concentration. Specifically, the mean oscillation frequency increased from 1.4±0.49 min−1 for 1mM caffeine, to 3.5±1.1 min−1 for 3mM caffeine (95% confidence interval), the increase between the two means being statistically significant at the 1% level. This is consistent with an increase in RyR sensitization as shown by Figs. 2A-B. We note however substantial variation in the Ca2+ response, shape and frequency between cells (Figure E5). A similarly large cell variability is observed for KCl-induced Ca2+ oscillations (Figure E6). By contrast, stimulation of ASMCs with higher concentrations of caffeine (>6mM caffeine; 10 mM shown in Fig. 2C(iii)) induced a steady elevated [Ca2+ ]i . These results are explained by the model predictions for luminal sensitization of the RyR (Fig. 2B(iii)), and are similar to the prior results obtained with 20mM caffeine [1]. In summary, these experimental results and model predictions indicate that RyR sensitization can lead to significant changes in the Ca2+ dynamics of ASMCs in resting conditions.

3.2

Effect of RyR sensitization on agonist-induced Ca2+ oscillations

Because ASMCs utilize Ca2+ oscillations to induce contraction and this contraction may be related to airway hypersensitivity, we use our model to explore how RyR sensitivity may alter agonistinduced Ca2+ oscillations. There is evidence that [Ca2+]SR is reduced during agonist-induced Ca2+ oscillations ([4] ; Figure E1A). We therefore expect that cytosolic sensitization of the RyR, if not accompanied by substantial luminal sensitization, will have negligible effect on agonist-induced Ca2+ oscillations. This is because in the absence of luminal sensitization, the opening probability of RyR decreases quickly with decreasing [Ca2+ ]SR . As a consequence, we begin our study of the effect of RyR sensitization on agonist-induced Ca2+ oscillations by assuming luminal RyR sensitization only, and then discuss how these results are modified if there is concomitant cytosolic RyR sensitization. We simulate pre-existing luminal RyR sensitization by setting rs to specific values (Figs. 3A and B) and mimic agonist application by increasing the IP3 concentration (p) from zero to either p = 0.5 (Fig. 3A) or p = 1.5 (Fig. 3B) at t = 200s. These simulations show that low or moderate luminal RyR sensitization has little effect on Ca2+ oscillations induced by low IP3 concentrations (p = 0.5); a small effect on Ca2+ oscillation frequency is noticeable with a RyR sensitivity rs = 3 (compare middle and bottom panels in Fig. 3A; see also red curve in Fig 3C).



By contrast, moderate luminal RyR sensitization (rs = 3) has a greater effect on Ca2+ oscillations induced by higher IP3 concentration (p = 1.5, Fig 3B). The Ca2+ oscillations are abolished and replaced with sustained elevated Ca2+ (blue trace). With higher levels of luminal RyR sensitization (rs = 4.25), the substantially elevated [Ca2+ ]i that follows the few initial slow Ca2+ oscillations is not significantly affected by an increase in [IP3 ] (green trace). Unfortunately, our experimental approach to test these predictions of RyR sensitisation effect on agonist-induced Ca2+ oscillations with caffeine were confounded by the inhibitory action of caffeine on the IP3 R [18, 19] (see also Discussion). We found the frequency of agonist-induced Ca2+ oscillations to be smaller (instead of larger) in the presence of caffeine than in its absence (Figs. E7-E9), even at large agonist concentrations ([MCh]=800nM). This outcome can be reproduced with the model by assuming that caffeine inhibits the IP3 R (e.g., decreases IP3 R conductance) in addition to increasing luminal RyR sensitization (simulations not shown). A summary of the interaction between luminal RyR sensitization and agonist stimulation is shown in Figures 3C and D. The frequency of Ca2+ oscillations (red curve) at low or high IP3 concentrations is relatively constant for increasing RyR sensitivity up to rs 2.5. The same applies for mean [Ca2+]i (dashed black curve). However, at higher RyR sensitivities (rs > 3), the mean [Ca2+]i increases gradually with increasing rs and Ca2+ oscillations are suppressed. Figure 4 shows how Fig. 3A (middle panel) and Fig. 3C are modified if we assume cytosolic RyR sensitization (rc = 0.5 or rc = 1) in addition to luminal RyR sensitization (rs = 3 in Fig. 4A). Moderate cytosolic sensitization (rc = 0.5) increases the frequency of agonist-induced Ca2+ oscillations (Fig. 4A, cyan trace) while large cytosolic sensitization (rc = 1) suppresses the Ca2+ oscillations (Fig 4A, magenta trace). Fig. 4B-C shows that cytosolic sensitization reduces the range of luminal sensitization (rs ) in which agonist-induced Ca2+ oscillations persist, and renders the Ca2+ oscillation frequency sensitive to luminal sensitization (compare with Fig. 3C). Fig. 5 underscores the important implications of our model findings by showing how the mean [Ca2+]i as a function of [IP3 ] is modified by increasing RyR sensitization. For any given agonist concentration (dotted vertical lines represent p = 0.5 and 1.5 that are used in Fig. 3), the mean [Ca2+]i is always higher in the presence of RyR sensitization; that is the greater the RyR sensitization, the greater the mean [Ca2+]i . These response curves also show that the range of IP3 concentrations that can induce Ca2+ oscillations (dashed lines) shrinks and disappears as luminal RyR sensitization is increased. In particular, the overlap between Ca2+ oscillations for rs = 0 (dashed blue curve) and steady Ca2+ for rs = 3 (solid red curve) over a large range of [IP3 ] (p = 0.8 to 2.3 for rc = 0) implies


Page 8 of 36


Page 9 of 36

that moderate changes in RyR sensitivity can have a large impact on the dynamics of the ASMC Ca2+ signaling. In summary, increasing luminal RyR sensitization progressively impairs agonistinduced Ca2+ oscillations and increases mean [Ca2+ ]i . A simultaneous increase in cytosolic RyR sensitization accelerates the suppression of the Ca2+ oscillations but does not induce an additional increase in mean [Ca2+ ]i at large IP3 concentrations where [Ca2+ ]i is already steady (compare blue and magenta curves).

4

Discussion

In this work, we have investigated the effect of RyR sensitization on Ca2+ dynamics in ASMC, using a dual approach of mathematical modeling and experimental observation of ASMCs in mouse lung slices. Our objective was to use our model to explore the hypothesis that an increase in RyR sensitivity, which may result from on-going airway inflammation, will alter the basic Ca2+ dynamics of ASMCs and their responses to contractile agonists. Any changes in the Ca2+ dynamics of ASMCs would be expected to have consequences for ASMC proliferation and airway hyper-responsiveness that are typical of asthma. A key advance in our model of the Ca2+ dynamics of ASMCs that allows us to explore the effect of RyR sensitization is the incorporation of SOCE and the interaction of the luminal Ca2+ concentration of the SR with the RyR. The regulation of SOCE is determined by the [Ca2+ ]SR which in turn is highly dependent on the open probability of the RyR which is itself dependent on [Ca2+ ]SR . Agonists that produce IP3 cause Ca2+ release from the SR, a gradual decrease of SR Ca2+ concentration, and gradual buildup of SOCE. KCl, on the other hand, causes increased Ca2+ influx, overloading of the SR, and release of Ca2+ through activated RyR. In this latter case, the SR is overloaded and so SOCE does not occur; even the Ca2+ spikes do not deplete the SR enough to cause significant SOCE. Caffeine, however, enables RyR activation in the absence of SR overloading. Hence, the Ca2+ spikes induced by caffeine may deplete the SR sufficiently to result in increased SOCE. Thus, although caffeine and KCl both induce RyR opening, they have quite different effects on SOCE. It is also important to note that, in our model, SOCE acts like a low-pass filter, responding not to each individual Ca2+ spike, but to mean levels of SR depletion or overfilling. Thus, individual Ca2+ spikes do not activate spikes of Ca2+ entry through SOC. The idea that the RyR may be involved in on-going inflammatory processes and may influence airway responsiveness has been previously suggested. For example, the cytokines TNF-α and IL-



13 were reported to increase the [Ca2+ ]i responses of ASMC to contractile agonists; a mechanism that correlates with activation of CD38, the production of cyclic ADP-ribose (cADPr) and the modification of the RyR [12, 8]. Experimental results obtained with beta-escin permeabilized ASMCs suggest that cADPr has a direct effect on the RyR [13].

4.1

RyR sensitization stimulates slow Ca2+ oscillations

The first prediction of the model, that is experimentally confirmed, is that small increases in cytosolic and/or luminal RyR sensitivity result in low-frequency Ca2+ oscillations. Although the consequences of these Ca2+ oscillations for ASMC physiology are not currently known, we propose two possibilities. The first and most obvious would be an increase in ASMC tone. However, in mouse airways, low-frequency Ca2+ oscillations do not induce substantial contraction; each ASMC is observed to twitch asynchronously. This is a result of the combination of a low sensitivity of mouse ASMCs to Ca2+ and the short duration of myosin light chain kinase activation. By contrast, other species including human, show greater contraction to low frequency Ca2+ oscillations. The second, but less obvious, consequence of low frequency oscillations could be a change in gene expression. In previous studies, Ca2+ oscillations were generated with the repetitive photolytic release of IP3 [20] or the alternate application of calcium and calcium-free solutions [21] to demonstrate that gene expression was initiated by low frequency Ca2+ oscillations. Ca2+ oscillations generated by IP3 at 1/min stimulated more gene expression than at 0.5/min or 2/min or by a sustained plateau [20]. The effectiveness of Ca2+ oscillations with a period of >100 seconds declined when using alternating Ca2+ solutions [21]. In a similar manner, Ca2+ oscillations may also affect enzyme activity as part of a frequency-modulated (FM) control system [22]. Thus, the stimulation of Ca2+ oscillations may have significant effects on the basic phenotype of the ASMCs. The idea that slow Ca2+ oscillations are associated with asthma is supported by the finding of spontaneous Ca2+ oscillations in human ASMCs from biopsies [23].

4.2

RyR sensitization increases mean [Ca2+ ]i

A second major prediction of the model is that moderate or large luminal RyR sensitization results in higher mean (or elevated) [Ca2+ ]i when the ASMCs are exposed to any concentration of contractile agonist. This prediction is consistent irrespective of whether luminal RyR sensitization is moderate and does not abolish the agonist-induced Ca2+ oscillations, or whether the RyR sensitization is high and converts the Ca2+ oscillations into a sustained Ca2+ elevation. Cytosolic RyR sensitization


Page 10 of 36


Page 11 of 36

can enhance the effect of luminal RyR sensitization on mean [Ca2+ ]i when luminal sensitization is moderate and not capable by itself to abolish agonist-induced Ca2+ oscillations. The immediate implication of a higher mean [Ca2+ ]i is that, in the presence of agonist, RyR sensitization could generate greater ASMC contraction. This, in turn, suggests a possible mechanism for airway hyperresponsiveness (AHR). The concept that increased (sustained) mean [Ca2+ ]i results in increased contraction is demonstrated in Ca2+ permeabilized human lung slices [2] where, for a given agonist concentration, airway contraction increases with constant [Ca2+ ]i imposed. Thus, with RyR sensitization, the mean [Ca2+ ]i would be higher than normal resulting in additional contraction. A second potential consequence of the replacement of agonist-induced Ca2+ oscillations with a sustained elevation in Ca2+ is stimulation of Ca2+ -dependent mechanisms that have slow response kinetics. It is already well known that the frequency of Ca2+ oscillation, and the shape of the oscillation, are important controllers of contraction [24]. However, there are potentially many other ways in which the frequency of the Ca2+ oscillation can control cellular processes; in the presence of RyR sensitization, recurrent exposure to agonist is likely to stimulate the ASMC in ways other than simple contraction. Unfortunately, however, we were unable to test this second prediction. Although caffeine is known to increase the sensitivity of the RyR to luminal Ca2+ , it also inhibits the IP3 R. It is thus not possible to test the effects of RyR sensitization in the absence of other confounding effects. Experimental results are shown in Figures E7–E9 of the Supplementary Material. Addition of MCh and caffeine results in a plateau of increased Ca2+ , but this plateau has superimposed slow oscillations, a result that does not agree with our model predictions. It is possible to modify our model to include the potential effects of caffeine inhibition of the IP3 R, and simulations with such a modified model display much better agreement with the experimental results shown in Figures E7–E9, but such modifications are speculative and outside the scope of the study here. It is left to future work to develop more rigorously a model of caffeine inhibition of the IP3 R. Given the impossibility of testing the effect of RyR sensitisation on agonist-induced Ca2+ oscillations with caffeine, due to its inhibitory effect on the IP3 R, we have looked for other drugs which could possibly sensitize the RyR without inhibiting the IP3 R. We found that pentifylline induces slow Ca2+ oscillations similar to those induced by caffeine (Figs. E10–E11 in Supplementary Material), but it is not clear whether pentifylline does so via luminal sensitization of the RyR (which is necessary to prevent RyR inhibition during agonist-induced Ca2+ oscillations) or via another mechanism.



However, like caffeine, pentifylline impairs agonist-induced Ca2+ oscillations (Figs. E12- E13): this could be due to an inhibition of the IP3 R, similar to that induced by caffeine, or possibly to the absence of luminal RyR sensitization. We have also tried the drug Chlorocresol [25] but it did not induce Ca2+ oscillations.

4.3 Limitations of the model In this work, we have modeled ASMC Ca2+ dynamics in a deterministic and homogenous manner although Ca2+ channels operate stochastically, and are not distributed homogeneously on the SR membrane. The extent to which this stochasticity translates into “macroscopic” (cell-level) Ca2+ dynamics depends on the cell type, the Ca2+ channels involved and the level of stimulation (e.g., agonist concentration). The spatial distribution of Ca2+ channels on the SR membrane could also be modified in disease. In ASMCs, as well as in other cell types, the distribution of the occurrence of Ca2+ spikes during agonist-induced Ca2+ oscillations is well described by a Poisson process, characterized by a linear dependence between the standard deviation and the mean of the interspike interval (ISI) [26]. However, the main properties of these stochastic Ca2+ oscillations can be predicted by a deterministic mathematical model [27]. Hence, the stochastic nature of the process does not preclude the value of a deterministic description of Ca2+ dynamics as used in this study. The mathematical model used here was originally designed to account for agonist-induced Ca2+ oscillations and SOCE in ASMC of human lung slices at room temperature. These Ca2+ oscillations are much slower than Ca2+ oscillations in mouse (10/min vs. 30/min). The model is therefore qualitative in this respect. Similarly, the RyR-mediated Ca2+ oscillations have a profile that is not fully reproduced by the model. Some differences in the Ca2+ profile of the model and data are likely the result of the affinity of Oregon Green BAPTA-1-AM that was used to report changes in Ca2+ (Kd ∼ 0.2µM). This reporter is saturated at Ca2+ above (1 µM). Hence, peak amplitudes of Ca2+ oscillations are unlikely to be reproduced in the fluorescent signal. Similarly, the kinetics of OGB may not be fast enough to track the rapid changes in [Ca2+ ]i . This could explain the elevated fluorescent baseline observed during agonist-induced Ca2+ oscillations, which is not apparent during RyR-mediated Ca2+ oscillations. However, our model accounts for a key feature of Ca2+ oscillation frequencies in ASMC; that is IP3 R-mediated Ca2+ oscillations are faster than RyR-mediated Ca2+ oscillations. This is because IP3 R gating is mainly governed by inhibition of the receptor by [Ca2+ ]i , whereas RyR gating is essentially governed by the level of Ca2+ store depletion. Since the refilling of the SR with Ca2+


Page 12 of 36


Page 13 of 36

occurs on a longer timescale than recovery of IP3 R from inhibition, RyR-mediated oscillations have a lower frequency than IP3 R-mediated oscillations. The frequencies of these two types of oscillations are closer in human ASMC (IP3 R; 10/min vs. RyR; 3/min) than in mouse ASMC (IP3 R; 30/min vs. RyR; 3/min), which suggests the effect of RyR sensitization on agonist-induced Ca2+ oscillations could be larger in human ASMCs (since oscillations of closer frequency are more likely to interact).

4.4 Scope of the experimental results As mentioned above, we have used mouse ASMC in this study. These exhibit faster IP3 R-mediated Ca2+ oscillations than human ASMC upon agonist stimulation, but both species exhibit maximum contraction (about 55% in both cases) at the highest end of the frequency spectrum (e.g., Fig.6 in [28]). Hence, the two species respond differently to a given Ca2+ oscillation frequency. Our predictions are however independent of this quantitative difference since in both cases, we expect RyR-sensitization-induced Ca2+ oscillations to enhance agonist-induced Ca2+ oscillations; the size of the effect will of course depend on the species, but the prediction is qualitatively the same. The key observation that increased RyR sensitivity can underlie unexpected complications in ASMC Ca2+ signaling and downstream physiology suggests the RyR as a target for therapeutic drugs. However, the exact nature of such a drug is difficult to predict since the mechanism of luminal Ca2+ sensitivity is not well understood. Compounds that could prevent Ca2+ binding to the RyR would serve to reduce its Ca2+ sensitivity. This might take the form of a Ca2+ buffer within the SR; the protein calsequestrin serves such a process in the SR. Since the SR Ca2+ content is a key parameter, drugs reducing SERCA pump activity or SOCE may also counteract increased RyR sensitivity by helping to maintain reduced SR Ca2+ content.

Acknowledgments BSB and HC were supported by funding from the Medical Research Council (New Investigator Grant G0901174). JS and MJS were supported by the NIH (grant HL103405).

References [1] Perez, J. F. and Sanderson, M. J. The frequency of calcium oscillations induced by 5-HT, ACH, and KCl determine the contraction of smooth muscle cells of intrapulmonary bronchioles. J Gen Physiol 2005; 125(6):535–53.



[2] Ressmeyer, A.-R., Bai, Y., Delmotte, P. F., Uy, K. F., Thistlethwaite, P., Fraire, A., Sato, O., Ikebe, M., and Sanderson, M. J. Human airway contraction and formoterol-induced relaxation is determined by Ca2+ oscillations and Ca2+ sensitivity. Am J Respir Cell Mol Biol 2010; 43(2):179–91. [3] Bai, Y. and Sanderson, M. J. The contribution of Ca2+ signaling and Ca2+ sensitivity to the regulation of airway smooth muscle contraction is different in rats and mice. Am J Physiol Lung Cell Mol Physiol 2009; 296(6):L947–58. [4] Wang, I. Y., Bai, Y., Sanderson, M. J., and Sneyd, J. A mathematical analysis of agonist- and KCl-induced Ca2+ oscillations in mouse airway smooth muscle cells. Biophys J 2010; 98(7):1170– 81. [5] Trian, T., Benard, G., Begueret, H., Rossignol, R., Girodet, P.-O., Ghosh, D., Ousova, O., Vernejoux, J.-M., Marthan, R., Tunon-deLara, J.-M., and Berger, P. Bronchial smooth muscle remodeling involves calcium-dependent enhanced mitochondrial biogenesis in asthma J Exp Med 2007; 204(13):3173–81. [6] White, T. A., Xue, A., Chini, E. N., Thompson, M., Sieck, G. C., and Wylam, M. E. Role of transient receptor potential C3 in TNF-alpha-enhanced calcium influx in human airway myocytes. Am J Respir Cell Mol Biol 2006; 35(2):243–51. [7] Moynihan, B., Tolloczko, B., Michoud, M.-C., Tamaoka, M., Ferraro, P., and Martin, J. G. MAP kinases mediate interleukin-13 effects on calcium signaling in human airway smooth muscle cells. Am J Physiol Lung Cell Mol Physiol 2008; 295(1):L171–7. [8] Sieck, G. C., White, T. A., Thompson, M. A., Pabelick, C. M., Wylam, M. E., and Prakash, Y. S. Regulation of store-operated Ca2+ entry by CD38 in human airway smooth muscle. Am J Physiol Lung Cell Mol Physiol 2008; 294(2):L378–85. [9] Mahn, K., Hirst, S. J., Ying, S., Holt, M. R., Lavender, P., Ojo, O. O., Siew, L., Simcock, D. E., McVicker, C. G., Kanabar, V., Snetkov, V. a., O’Connor, B. J., Karner, C., Cousins, D. J., Macedo, P., Chung, K. F., Corrigan, C. J., Ward, J. P. T., and Lee, T. H. Diminished sarco/endoplasmic reticulum Ca2+ ATPase (SERCA) expression contributes to airway remodelling in bronchial asthma. PNAS 2009; 106(26):10775–80.


Page 14 of 36


Page 15 of 36

[10] Sathish, V., Thompson, M. A., Bailey, J. P., Pabelick, C. M., Prakash, Y. S., and Sieck, G. C. Effect of proinflammatory cytokines on regulation of sarcoplasmic reticulum Ca2+ reuptake in human airway smooth muscle. Am J Physiol Lung Cell Mol Physiol 2009; 297(1):L26–34. [11] Sathish, V., Delmotte, P. F., Thompson, M. a., Pabelick, C. M., Sieck, G. C., and Prakash, Y. S. Sodium-calcium exchange in intracellular calcium handling of human airway smooth muscle. PloS One 2011; 6(8):e23662. [12] Deshpande, D. a., Dogan, S., Walseth, T. F., Miller, S. M., Amrani, Y., Panettieri, R. a., and Kannan, M. S. Modulation of calcium signaling by interleukin-13 in human airway smooth muscle: role of CD38/cyclic adenosine diphosphate ribose pathway. Am J Respir Cell Mol Biol 2004; 31(1):36–42. [13] Prakash, Y. and Kannan, M. Role of cyclic ADP-ribose in the regulation of [Ca2+ ]i in porcine tracheal smooth muscle American Journal of Physiology - Cell Physiology (1998) 274:C1653– C1660. [14] Jude, J., Dileepan, M., Panettieri, R., Walseth, T. F., and Kannan, M. S. Altered CD38/cyclic ADP-ribose signaling contributes to the asthmatic phenotype. J Allergy 2012; 2012:289468. [15] Kong, H., Jones, P. P., Koop, A., Zhang, L., Duff, H. J., and Chen, S. R. W. Caffeine induces Ca2+ release by reducing the threshold for luminal Ca2+ activation of the ryanodine receptor. The Biochemical journal 2008; 414(3):441–52. [16] Porta, M., Zima, A. V., Nani, A., Diaz-Sylvester, P. L., Copello, J. A., Ramos-Franco, J., Blatter, L. A., and Fill, M. Single ryanodine receptor channel basis of caffeine’s action on Ca2+ sparks. Biophys J 2011; 100(4):931–8. [17] Croisier, H., Tan, X., Perez-Zoghbi, J. F., Sanderson, M. J., Sneyd, J., and Brook, B. S. Activation of store-operated calcium entry in airway smooth muscle cells: insight from a mathematical model. PloS One 2013; 8(7):e69598. [18] Parker, I. and Ivorra, I. Caffeine inhibits inositol trisphosphate-mediated liberation of intracellular calcium in Xenopus oocytes J Physiol 1991; 433:229–40. [19] Bezprozvanny, I., Bezprozvannaya, S., and Ehrlich, B. E. Caffeine-induced inhibition of inositol(1,4,5)-trisphosphate-gated calcium channels from cerebellum Mol Biol Cell 1994; 5(1):97–103. 15 Copyright © 2015 by the American Thoracic Society


[20] Li, W., Llopis, J., Whitney, M., Zlokarnik, G., and Tsien, R. Y. Cell-permeant caged InsP3 ester shows that Ca2+ spike frequency can optimize gene expression Nature 1998; 392(6679):936–41. [21] Dolmetsch, R. E., Xu, K., and Lewis, R. S. Calcium oscillations increase the efficiency and specificity of gene expression Nature 1998; 392(6679):933–6. [22] Berridge, M. J. The AM and FM of calcium signalling Nature 1997; 386(6627):759–60. [23] Sweeney, D., Hollins, F., Gomez, E., Saunders, R., Challiss, R. A. J., and Brightling, C. E. [Ca2+ ]i oscillations in ASM: relationship with persistent airflow obstruction in asthma Respirology 2014; 19(5):763–6. [24] Roux, E., Guibert, C., Savineau, J. P., and Marthan, R. [Ca2+ ]i oscillations induced by muscarinic stimulation in airway smooth muscle cells: receptor subtypes and correlation with the mechanical activity. Br J Pharmacol 1997; 120(7):1294–301. [25] Al-Mousa, F. and Michelangeli, F. Commonly used ryanodine receptor activator, 4-chloro-mcresol (4CmC), is also an inhibitor of SERCA Ca2+ pumps Pharmacol Rep (2009) 61(5):838–42. [26] Skupin, A., Kettenmann, H., Winkler, U., Wartenberg, M., Sauer, H., Tovey, S. C., Taylor, C. W., and Falcke, M. How does intracellular Ca2+ oscillate: by chance or by the clock? Biophys J 2008; 94(6):2404–11. [27] Cao, P., Tan, X., Donovan, G., Sanderson, M. J., and Sneyd, J. A deterministic model predicts the properties of stochastic calcium oscillations in airway smooth muscle cells. PLoS Comput Biol 2014; 10(8):e1003783. [28] Lauzon, A.-M., Bates, J. H. T., Donovan, G., Tawhai, M., Sneyd, J., and Sanderson, M. J. A multi-scale approach to airway hyperresponsiveness: from molecule to organ. Front Physiol 2012; 3:191(1–25). [29] Sneyd, J., Tsaneva-Atanasova, K., Reznikov, V., Bai, Y., Sanderson, M. J., and Yule, D. I. A method for determining the dependence of calcium oscillations on inositol trisphosphate oscillations. PNAS 2006; 103(6):1675–80. [30] De Young, G. W. and Keizer, J. A single-pool inositol 1,4,5-trisphosphate-receptor-based model for agonist-stimulated oscillations in Ca2+ concentration Proc Natl Acad Sci U S A 1992; 89(20):9895–9.


Page 16 of 36


Page 17 of 36

[31] Li, Y. and Rinzel, J. Equations for InsP3 receptor-mediated [Ca2+ ]i oscillations derived from a detailed kinetic model: a Hodgkin-Huxley like formalism J Theor Biol (1994) 166(4):461–473. [32] Tang, Y., Stephenson, J. L., and Othmer, H. G. Simplification and analysis of models of calcium dynamics based on IP3 -sensitive calcium channel kinetics. Biophys J 1996; 70(1):246–63. [33] Keener, J. and Sneyd, J. (2008) Mathematical Physiology, Springer-Verlag, New York second edition.



KCl

outside the cell

Ca2+

SOCC

methacholine

+

-

SERCA

PMCA G protein

GPCR

VOCC

IP 3

2+ Sarcoplasmic Reticulum

+

+

cel l cy to pl a sm

Ca

IP 3R +

+

CICR +

RyR

+

Ca2+

c aff ei n e

Figure 1: Schematic diagram of the major pathways involved in the model of Ca2+ dynamics in ASMCs. Solid black lines denote Ca2+ fluxes. Stimulation of cell-surface G protein-coupled receptors (GPCR) leads to production of IP3 and activation of the IP3 receptor (IP3 R). Subsequent release of Ca2+ from the SR leads to Ca2+ -induced Ca2+ release (CICR) through either IP3 R or ryanodine receptors (RyR). Depletion of the SR induces the opening of store-operated Ca2+ channels (SOCC) in the plasma membrane, but also inactivates RyR. Ca2+ is removed from the cytoplasm by plasma membrane ATPase (PMCA) or by SERCA. Addition of extracellular KCl leads to membrane depolarization and opening of voltage-operated Ca2+ channels (VOCC). The consequent increased Ca2+ influx causes (in the absence of agonist) overfilling of the SR and activation of the RyR. IP3 R and RyR interact via their effects on the concentration of Ca2+ in the cytosol and the SR. Here, we focus on the predicted effects of caffeine (and other compounds), which sensitizes the RyR to cytosolic and luminal Ca2+.


Page 18 of 36


Page 19 of 36

A – Cytosolic sensitization (i) r =0.75

2+

[Ca ] (µM)

c

(ii) r =1

(iii) r =1.75

c

c

2

2

2

1.6

1.6

1.6

1.2

1.2

1.2

0.8

0.8

0.8

0.4

0.4

0.4

0

0

0

100

200

300

0

100

200

time (s)

0

300

0

100

time (s)

200

300

time (s)

B – Luminal sensitization (i) r =3

2+

[Ca ] (µM)

s

(ii) r =3.5

(iii) r =4.25

s

s

2

2

2

1.6

1.6

1.6

1.2

1.2

1.2

0.8

0.8

0.8

0.4

0.4

0.4

0

0

0

100

200

300

0

100

time (s)

200

0

300

0

100

time (s)

200

300

time (s)

C – Experiments with caffeine (i) 1mM

(iii) 10mM

(ii) 3mM

2.5

2.5 3

2

F/F0

F/F0

F/F

0

2

1.5

1.5

1

1

1

1.5

0.5

50

100

150

200

250

2

0

300

50

100

time (s)

150

200

250

50

300

100

time (s)

150

200

250

300

time (s)

D – Luminal sensitization - store calcium (i) r =3

(ii) r =3.5

(iii) r =4.25

s

s

160

160

120

120

120

80

80

80

40

40

40

0

0

2+

[Ca ]

SR

(µM)

s 160

0

100

200

300

0 0

100

time (s)

200

300

0

100

time (s)

200

300

time (s)

E – Luminal sensitization - RyR and SOCC fraction of open channels

(ii) r =3.5

(i) r =3

s

(iii) r =4.25

s

s

1

1

1

0.8

0.8

0.8

0.6

0.6

0.6

0.4

0.4

0.4

0.2

0.2

0

0

100

200

time (s)

300

0

0.2 0

100

200

300

0

time (s)

0

100

200

300

time (s)

Figure 2: (A) Simulations of the effect of increasing levels of cytosolic RyR sensitization (rc ) on [Ca2+]i dynamics, starting from (rc , rs ) = (0, 1) (red dot in Fig. E3A). (B) Simulations of the effect of increasing levels of luminal RyR sensitization (rs ) on [Ca2+]i dynamics, starting from (rc , rs ) = (0, 0) (no sensitization, black dot in Fig. E3A). (C) Relative fluorescence signals indicating the [Ca2+ ]i responses induced by caffeine in ASMC of mouse lung slices (see Methods) for (i) 1mM caffeine, (ii) 3mM caffeine, and (iii) 10mM caffeine. (D) The predicted [Ca2+ ]SR associated with the [Ca2+]i dynamics modelled in (B). (E) The predicted fraction of open RyRs (magenta) and open SOCCs (blue) associated with the [Ca2+ ]i dynamics in (B).



A - - Low agonist

B - - High agonist p=0.5

p=1.5 2

r =4.25

[Ca ] (µM)

s

1

2+

2+

[Ca ] (µM)

2

0 100

200

300

0

100

200

300

400

2 p=0.5

rs=3

[Ca ] (µM)

1

2+

2+

[Ca ] (µM)

s

1

400

2

0

p=1.5

rs=3 1

0 0

100

200

300

400

0

2

100

200

300

400

2 p=0.5

rs=0

[Ca ] (µM)

1

2+

2+

r =4.25

0 0

[Ca ] (µM)

Page 20 of 36

0

p=1.5

rs=0 1

0 0

100

200 time (s)

300

400

0

200 time (s)

300

400

12 10

0.4

8 0.2

6 0

1

2

r

3

4

5

0.8

14 p=1.5

0.6

12 10

0.4

8 0.2 0

1

s

2

r

3

4

6 5

frequency (min−1)

0.6

i

14

p=0.5

frequency (min−1)

i

0.8

mean [Ca2+] (µ M)

D

mean [Ca2+] (µ M)

C

100

s

Figure 3: (A-B) Effect of agonist application on ASMC Ca2+ dynamics with different levels of pre-existing luminal RyR sensitization rs (increasing from top to bottom), in the absence of cytosolic RyR sensitization (rc = 0). (A) Low [IP3 ] (p = 0.5), and (B) high [IP3 ] (p = 1.5), applied at t = 200s. (C-D) Long-term mean [Ca2+ ]i (black) and oscillation frequency (red) as a function of luminal RyR sensitization for the two agonist levels in (A) and (B).



Page 21 of 36

1

0 0

100

200

0.6

12

0.4

10 8

0.2 0

c

1

2

3 r

1

0 200

2

400

p=0.5

r =3, r =1 s

300

i

100

2+

0 [Ca2+] (µM)

c

p=0.5

r =3, r =0.5 s

400

mean [Ca ] (µ M)

[Ca2+] (µM)

2

300

14

−1

c

p=0.5, r =0.5

c

1

0.8

200 time (s)

300

6

14

c

0.6

12 10

0.4

8 0.2

6 1

0 100

5

s

p=0.5, r =1

0 0

4

400

2

r

3

4

−1

s

0.8

frequency (min )

p=0. 5

r =3, r =0

2+ i

[Ca2+] (µM)

2

frequency (min )

B mean [Ca ] (µ M)

A

5

s

Figure 4: Effect of simultaneous cytosolic and luminal RyR sensitization on agonist-induced Ca2+ oscillations for low [IP3 ] (p = 0.5). (A) Increasing levels of cytosolic sensitization (from top to bottom) with RyR luminal sensitization rs = 3 (top panel is identical to middle panel in Fig 3A); (B-C) Long-term mean [Ca2+]i (black curves) and oscillation frequency (red curve) as a function of luminal RyR sensitization for the two non-zero sensitization levels used in (A) (compare with Fig. 3C).



Figure 5: Effect of different levels of RyR sensiti-

increasing sensitisation

0.6

zation on the entire “dose-response curve” relating mean [Ca2+ ]i to [IP3 ] during agonist stimulation. Dashed curves correspond to Ca2+ oscillations and solid curves to steady Ca2+ . Dotted vertical lines indicate the [IP3 ] values used in Fig. 3A and B. The bottom (red) curve is the dose-response curve in the absence of sensitization, hence it corresponds to the response curve in Fig. E2A. Grey curves represent unstable solutions.

i

mean [Ca2+] (µ M)

0.8

0.4 r =0, r =0 s

c

r =3, r =0

0.2

s

c

r =3, r =1 s

c

r =4.25, r =0 s

c

0 0

1

2

3

p


Page 22 of 36


Page 23 of 36

Supplemental material to: “Ryanodine receptor sensitization results in abnormal calcium signaling in airway smooth muscle cells” Huguette Croisier1 , Xiahui Tan2 , Jun Chen3 , James Sneyd4 , Michael J. Sanderson3 , Bindi S. Brook1 1 School of Mathematical Sciences, University of Nottingham, Nottingham, UK. 2 Lung Inflammation and Infection Laboratory, QIMR Berghofer Medical Research Institute, Herston, Queensland, Australia 3 Department of Microbiology and Physiological Systems, University of Massachusetts Medical School, Worcester, MA, USA. 4 Department of Mathematics, University of Auckland, Auckland, New Zealand.

E1 E1.1

Details of methods Lung slice preparation

Reagents: Unless specified, all reagents were obtained from Sigma Aldrich (St. Louis, MO, USA). Hanks’ buffered saline solution was supplemented with 20 mM HEPES (sHBSS) and adjusted to pH 7.4 with NaOH. Caffeine was prepared in sHBSS. Mouse housing conditions: BALB/c mice (Charles River Breeding Labs, Needham, MA, USA) were kept in 484 cm2 rectangular polysulfone cages containing Bed-o’ Cobs

1 4

inch bedding material

(The Andersons Inc, Maumee, OH, USA) which were autoclaved prior to use. Mice (4 per cage) were exposed to 12 hour day/night cycle, and provided with free access to irradiated Isopro 5P76 pellet diet (LabDiet, St Louis, MO, USA) and water (pH 2.8-3). All animal procedures were approved by the University of Massachusetts Medical School Institutional Animal Care and Use Committee (IACUC). Precision cut lung slice (PCLS) preparation: Female BALB/c mice (7-12 weeks old) were euthanized either by intraperitoneal injection of 0.3 mL of sodium pentabarbitone (50 mg/mL) or by cervical dislocation. After removal of the chest wall, the trachea was intubated with a 23G intravenous catheter for the inflation of the lungs with 1.3 ml of 1.8% warm agarose in sHBSS. Subsequently, a 0.3 ml volume of air was injected to expel the agarose from within the airways into the alveolar space. The agarose was gelled by cooling to 4degC with cold sHBSS. A vibratome (VF-

E1 Copyright © 2015 by the American Thoracic Society


Page 24 of 36

300, Precisionary Instruments, Greenville, NC, USA) was used to make ∼ 180 − 200 µm thick lung slices which were maintained in Dulbecco’s modified Eagle’s media (DMEM, Invitrogen, Carlsbad, CA) at 37degC in 10% CO2 /air. All experiments were conducted at body temperature (37degC) in a custom-made, temperature-controlled microscope chamber as described in [1].

E1.2

Full mathematical model

This work is based on our previous model [2], which accounts for the contribution of SOCE to Ca2+ dynamics of ASMC. It is extended here to incorporate the dynamics of RyR and voltage-operated Ca2+ channels (VOCC). The dynamics of c =[Ca2+ ]i and cs =[Ca2+ ]SR are governed by: dc = Jin − JP M CA + Jrel − JSERCA , dt dcs = γ(JSERCA − Jrel ), dt

(1)

Jin = JleakP M + JROCC + JSOCC + JV OCC ,

(2)

Jrel = (kIP R PIP R + kRyR PRyR + JSR )(cs − c).

(3)

where

The term JleakP M is a constant Ca2+ leak through unspecified plasma membrane (PM) channels, JROCC is the Ca2+ influx through receptor-operated Ca2+ channels (ROCC) and JSOCC the influx through SOCC:

Jleakin = α0 ,

(4a)

JROCC = α1 p,

(4b)

JSOCC (Pso ) = Vs Pso .

(4c)

where α0 and α1 are constants, p is the concentration of IP3 in the cytosol, Vs is the maximum SOCC flux, and Pso is the fraction of open SOCC. We assume, in accordance with available evidence [3], that [IP3 ] remains constant once the agonist concentration is fixed, so that p can also be considered as an indication of the agonist level. The evolution of Pso is modelled by



Page 25 of 36

dPso ∞ = (Pso (cs ) − Pso ) /τs , dt K4 ∞ Pso (cs ) = 4 s 4 . Ks + cs

(5a) (5b)

which accounts for the delay between Ca2+ store depletion and SOCE activation, with time constant τs and half-activation threshold Ks [2]. We use the expression from [4] to model the Ca2+ influx through VOCC (which occurs in response to membrane depolarisation by KCl): JV OCC (c) = −kV m2 ΦCa (c), m=

1 , 1 + e−(V −Vm )/km

ΦCa (c) =

V (c − co e− 1−e

2V F RT

F − 2V RT

(6) )

.

where m2 is the fraction of open VOCC at membrane potential V and ΦCa is the driving force for Ca2+ through the cell membrane (GHK equation). The other parameters are constants described in Table E1. The constants kIP R and kRyR in (3) are the maximum rates of Ca2+ flow through the IP3 R and RyR respectively. JSR is the constant rate of Ca2+ leak through the SR membrane. The IP3 R opening probability PIP R is modelled by [5, 6, 7]: PIP R (c, y) =

where

pc(1 − y) (p + K1 )(c + K5 )

3 ,

dy = Φ1 (c)(1 − y) − Φ2 y, dt (k42 K2 K1 + K4 p)c Φ1 (c) = k2− , K4 K2 (K1 + p) p + k42 K3 Φ2 = k2− . K3 + p

(7)

(8)

We also use the description from Wang et al. [4] for RyR dynamics:

PRyR (c, cs ) =

c4s c3 3 + c3 K 4 + c4 , KcR s sR

(9)

where KcR and KsR are the [Ca2+ ]i and [Ca2+ ]SR RyR half-activation thresholds respectively. This formulation implies that [Ca2+ ]i and [Ca2+ ]SR must simultaneously be large enough (i.e., compaE3 Copyright © 2015 by the American Thoracic Society


Page 26 of 36

rable respectively to KcR and KsR ) for RyR activation, and that the fraction of open RyR adapts instantaneously to changes in c and cs . Inactivation by (large) [Ca2+ ]i is neglected. Finally, the Ca2+ pumps are modelled using the expressions (e.g., [8]):

JP M CA (c) = Vp

c2 , Kp2 + c2

JSERCA (c) = Ve

c2 . Ke2 + c2

(10)

where Kp and Ke are the respective half-activation thresholds for PMCA and SERCA activation, and Vp and Ve the respective maximum Ca2+ rates. The full model writes: c2 dc =α0 + α1 p + Vs Pso − kV m2 ΦCa (c) − Vp 2 dt Kp + c2 c2 + (kIP R PIP R (c, y) + kRyR PRyR (c, cs ) + JSR )(cs − c) − Ve 2 , Ke + c2 dcs c2 =γ Ve 2 − (kIP R PIP R (c, y) + kRyR PRyR (c, cs ) + JSR )(cs − c) , dt Ke + c2 dy =Φ1 (c)(1 − y) − Φ2 y, dt dPso ∞ = (Pso (cs ) − Pso ) /τs . dt

(11)

We chose the model parameters KcR , KsR , kRyR , and kV in order to preserve the results obtained in [2] while accounting additionally for KCl-induced Ca2+ oscillations [4] (see details in next section). Based on this procedure, we obtained model parameters as given in Table E1. These parameter values define the unsensitized state from which we study the effect of RyR sensitization on Ca2+ dynamics in the absence and presence of agonist. Parameter name Max. RyR Ca2+ rate RyR [Ca2+ ]i activation threshold RyR [Ca2+ ]SR activation threshold VOCC conductance VOCC half-max. activation voltage VOCC activation curve steepness Resting potential Depolarising potential Extracellular [Ca2+ ] Faraday’s constant Gas constant Absolute temperature

symbol kRyR KcR0 KsR0 kV Vm km V V c0 F R T

value 0.333 0.25 250 0.833 -50 12 -70 -30 1300 96485 8345 310

units 1/s µM µM nS/µM mV mV mV mV µM C/mol mJ/(mol K) K

reference this work this work this work this work [4] [4] [4] [4] [9]

Table E1: Parameter values for RyR and VOCC Ca2+ fluxes (Eqs. (6) and (9)). See [2] for the other parameter values. This gives resting Ca2+ concentrations c = 0.0816µM and cs = 168µM.



Page 27 of 36

E1.3

Model parameter determination

The following rationales and observations guided our parameter choices: • KsR0 was chosen above the resting [Ca2+ ]SR because overfilling of the SR with Ca2+ is believed to be required for RyR opening [4]. As a consequence, upon agonist stimulation, RyR remain open only briefly since [Ca2+ ]SR decreases substantially below its resting value in the continued presence of agonist (Figure E1A). On the other hand, [Ca2+ ]SR remains large during KCl-induced Ca2+ oscillations (Figure E1B) because the RyR close quickly when [Ca2+ ]SR decreases, stopping the SR depletion from Ca2+ and allowing its refilling. Using smaller values of KsR0 than that retained increases both the amplitude of KCl-induced Ca2+ oscillations and the initial response to agonist stimulation. Indeed, the lower KsR , the more Ca2+ can be released by RyR before they are inactivated by the decreasing [Ca2+ ]SR ; • The cytosolic threshold KcR0 is set sufficiently above the resting [Ca2+ ]i so that the Ca2+ flux through RyR remains small at equilibrium. This ensures that the resting [Ca2+ ]i does not become too large, or the resting [Ca2+ ]SR too small, compared to the physiological values. At the same time, KcR0 cannot be too large either, because the larger it is, the slower RyRmediated Ca2+ oscillations, due to the greater time needed for [Ca2+ ]i to reach the RyR activation threshold. Concomitantly, the amplitude of the RyR-mediated Ca2+ oscillations becomes larger because the SR refills more with Ca between spikes. • The greater the maximum VOCC rate kV , the greater the frequency of KCl-induced Ca2+ oscillations, since the less time it takes to reach the necessary SR Ca2+ overfilling for spiking. The latency to first spike decreases concomitantly. • An increase in the RyR maximum Ca2+ rate kRyR induces a slight increase in the amplitude and in the frequency of KCl-induced Ca2+ oscillations • KcR0 , KsR0 , kRyR , and kV have to be such that the resting [Ca2+ ]i is '100nM and the resting [Ca2+ ]SR '200µM, while the frequency of agonist-induced oscillations frequency is of the order of 1-10/min and that of KCl-induced Ca oscillations of 1-3/min (human lung slice data [9, 2]) The results are illustrated in Figs. E1 and E2. Figs E1A and E1B show examples of agonistinduced (p = 1) and KCl-induced (V = −30) Ca2+ oscillations, respectively. Figs E2A and E2B show how the mean [Ca2+ ]i and the Ca2+ oscillation frequency vary as a function of [IP3 ] and membrane potential V, respectively. E5 Copyright © 2015 by the American Thoracic Society


Page 28 of 36

Although the amplitude of KCl-induced [Ca2+ ]i oscillations is larger than that of agonist-induced [Ca2+ ]i oscillations (compare Figures E1A and E1B), the mean [Ca2+ ]i (over an oscillation period) is smaller, due to the lower frequency. The reason the frequency curve for KCl-induced Ca2+ oscillations is biphasic is that the Ca2+ influx through VOCC is itself a biphasic function of membrane potential (Eq. 6). A

B p=1

V=−30mV 1

1

c cs/100

P

PSOC

SOC

0.5

0.8

0.4

RyR

0.5

0.8

open

PRyR

P

1.2

P

PIPR

[Ca2+] (µM)

1.2

0

c cs/100

1.6

Popen

[Ca2+] (µM)

1.6

0.4

0

40

80

120

0

160

0

0

100

t (s)

200

0 300

t (s)

Figure E1: (A) Agonist-induced Ca2+ oscillations (p = 1, V = −70mV), and (B) KCl-induced Ca2+ oscillations (p = 0, V = −30mV) obtained with the parameter values given in Table E1. Note that for the agonist-induced oscillations the SR is depleted and the SOCC current is large. However, for KCl-induced oscillations, the SR is overfilled and there is no SOCC current. This contrasts with the model simulations shown in Fig. 2E, where caffeine-induced sensitization of the RyR leads to Ca2+ oscillations even when the SR is (partially) depleted. In that case a larger SOCC current is predicted.

A

B 1

1

4

6 0.2 0 0

0.5

1

1.5

2

2 2.5

3 0.6 2 0.4 1 0.2 0 −50

p

0

−1

0.8

frequency (min )

0.4

2+

10

mean [Ca ]i (µ M)

0.6

frequency (min−1)

mean [Ca2+]i (µ M)

14 0.8

0 50

p

Figure E2: Average [Ca2+ ]i (black) and Ca2+ oscillation frequency (red) as a function of (A) IP3 concentration p, and (B) membrane potential V . Solid curves represent steady Ca2+ and dashed curves Ca2+ oscillations. Stable (respectively unstable) solutions are shown as thick (respectively thin) lines.



Page 29 of 36

E2

The modelling results in greater detail

Figure E3 summarises and extends the findings of Figures 2A and B. Figure E3A shows the parameter space (i.e., the values of rc and rs ) for which sustained Ca2+ oscillations exist, as indicated by the space contained within the blue curve. From its shape, we can deduce that the range of luminal sensitization (rs , vertical axis) in which Ca2+ oscillations occur shortens as rc is increased (and disappear for rc > 1.2). Figures E3B and C show the predicted (long-term) mean [Ca2+ ]i and oscillation frequency as a function of rc (for rs =1) and rs (for rc =0), respectively. For rc < 0.375 in Figure E3B and rs < 1.55 in Figure E3C, there are no Ca2+ oscillations (solid curves) and [Ca2+ ]i remains close to the resting level (or returns to it after a transient increase). Ca2+ oscillations occur for 0.375 ≤ rc ≤ 1.375 and 1.55 ≤ rs ≤ 3.9 (dashed curves). The frequency of these RyR-mediated Ca2+ oscillations is substantially lower than that of agonist-induced Ca2+ oscillations, and similar to the frequency of KCl-induced Ca2+ oscillations that are also mediated via the RyR (cf. Figs. E1 and E2). Ca2+ oscillations cease at rc ∼ 1.5 and rs ∼ 4, that is where KcR ∼ Ke and KsR ∼ Ks , respectively. We can understand that these are critical values since (i) Ke is close to the resting [Ca2+ ]i (cf. Table E1), so that KcR ∼ Ke implies that the fraction of open RyRs becomes virtually insensitive to [Ca2+ ]i , and (ii) KsR ∼ Ks implies that RyRs remain open at [Ca2+ ]SR levels where SOCE is substantially activated, ensuring that [Ca2+ ]SR does not become low enough for RyRs to close (which would allow a new cycle of refilling of the SR with Ca2+ , SOCE inactivation, RyR opening, etc).

oscill.

2 1

6 0.6 4 0.4 2 0.2

2+

0.8

1

mean [Ca ]i (µ M)

3

8

−1

no oscill.

mean [Ca2+]i (µ M)

4

rs

C 1

6

0.8 4 0.6 0.4 2 0.2

frequency (min−1)

B 5

frequency (min )

A

0 −2

−1

0

1

rc

2

0 0

0.5

1

1.5

2

0 2.5

rc

0 0

1

2

3

4

0 5

rs

Figure E3: (A) Region in the (rc ,rs ) plane in which stable RyR-mediated oscillations exist (the plot is restricted to rc ≤ 2.5 and rs ≤ 5 because KcR < 0 for rc > 2.5, and KsR < 0 for rs > 5). The black dot, where (rc , rs ) = (0, 0) represents the unsensitized case (KcR = KcR0 and KsR = KsR0 ). The red dot, where (rc , rs ) = (0, 1), is the starting point used to study the effect of increasing cytosolic sensitization (increasing rc only from (rc , rs ) = (0, 0) does not induce Ca2+ oscillations). (B-C) Effect on average [Ca2+ ]i (black) and Ca2+ oscillation frequency (red) of (B) cytosolic and (C) luminal RyR sensitization, following the paths in the parameter plane shown by the dotted lines in (A). The dotted vertical lines in (B) and (C) indicate the parameter values used in the simulations of Fig. 2. Thin grey curves represent unstable solutions (which cannot be observed experimentally).



E3 E3.1

Page 30 of 36

The experimental results in greater detail The caffeine dose-response curve

Mean frequency of oscillation (min-1)

5 4.5 n=20 4 n=2

3.5 3 n=8 2.5 2 n=15 1.5 1 0.5

1

1.5

2

2.5

3

3.5

4

4.5

Caffeine concentration (mM)

Figure E4: Mean oscillation frequency as a function of caffeine concentration (in the absence of agonist stimulation). Bars represent standard error of the mean (SEM). The number of experimental results for 4mM caffeine is too small to give useful statistics (e.g., 95% confidence interval=mean±12.7SEM).



Page 31 of 36

E3.2

Variability in RyR-mediated Ca2+ oscillations

A

B 2

F/F0

F/F0

2 1.5

1.5

1

1 50

100

150

200

250

300

F/F0

F/F0

2 1.5

150

200

250

300

50

100

150

200

250

300

50

100

150

200

250

300

50

100

150

200

250

300

50

100

150

200

250

300

2

1 50

100

150

200

250

300

2.5

4

2

F/F0

F/F0

100

1.5

1

1.5 1

3 2 1

50

100

150

200

250

300

2.5

2

F/F0

F/F0

50

2.5

2 1.5 1

1.5 1 0.5

50

100

150

200

250

300

2.5

F/F0

F/F0

3 2

2 1.5 1

1 50

100

150

200

250

300

time(s)

time(s)

Figure E5: Fluorescence traces of [Ca2+ ]i responses induced by (A) 1mM and (B) 3mM caffeine in ASMC in mouse lung slices.

F/F0

3 2 1 50

100

150

200

250

300

50

100

150

200

250

300

50

100

150

200

250

300

50

100

150

200

250

300

50

100

150

200

250

300

F/F0

2 1.5 1

F/F0

3 2 1

F/F0

2 1.5

F/F0

1

2 1.5 1

time(s)

Figure E6: Fluorescence traces of [Ca2+ ]i responses induced by 100mM KCl in ASMC in mouse lung slices.



E3.3

Page 32 of 36

Caffeine effect on MCh-induced Ca2+ oscillations

Figure E7: Fluorescence traces of the effect of MCh on Ca2+ dynamics in the absence and in the presence of 2mM caffeine ([MCh]=100nM).


Page 33 of 36






E3.4

Page 34 of 36

Pentifylline-induced Ca2+ dynamics

Figure E10: Fluorescence traces of [Ca2+ ]i responses induced by 200µM pentifyllline in ASMC in mouse lung slices.

Figure E11: Fluorescence traces of [Ca2+ ]i responses induced by 1mM pentifyllline in ASMC in mouse lung slices.



Page 35 of 36

E3.5

Pentifylline effect on MCh-induced Ca2+ oscillations

Figure E12: Fluorescence traces of the [Ca2+ ]i response to 100nM MCh in the absence and in the presence of 300µM pentifylline.

Figure E13: Fluorescence traces of the [Ca2+ ]i response to 800nM MCh in the absence and in the presence of 300µM pentifylline.



Page 36 of 36

References [1] Bai, Y. and Sanderson, M. J. The contribution of Ca2+ signaling and Ca2+ sensitivity to the regulation of airway smooth muscle contraction is different in rats and mice. Am J Physiol Lung Cell Mol Physiol 2009; 296(6):L947–58. [2] Croisier, H., Tan, X., Perez-Zoghbi, J. F., Sanderson, M. J., Sneyd, J., and Brook, B. S. Activation of store-operated calcium entry in airway smooth muscle cells: insight from a mathematical model. PloS One 2013; 8(7):e69598. [3] Sneyd, J., Tsaneva-Atanasova, K., Reznikov, V., Bai, Y., Sanderson, M. J., and Yule, D. I. A method for determining the dependence of calcium oscillations on inositol trisphosphate oscillations. PNAS 2006; 103(6):1675–80. [4] Wang, I. Y., Bai, Y., Sanderson, M. J., and Sneyd, J. A mathematical analysis of agonist- and KCl-induced Ca2+ oscillations in mouse airway smooth muscle cells. Biophys J 2010; 98(7):1170– 81. [5] De Young, G. W. and Keizer, J. A single-pool inositol 1,4,5-trisphosphate-receptor-based model for agonist-stimulated oscillations in Ca2+ concentration Proc Natl Acad Sci U S A 1992; 89(20):9895–9. [6] Li, Y. and Rinzel, J. Equations for InsP3 receptor-mediated [Ca2+ ]i oscillations derived from a detailed kinetic model: a Hodgkin-Huxley like formalism J Theor Biol (1994) 166(4):461–473. [7] Tang, Y., Stephenson, J. L., and Othmer, H. G. Simplification and analysis of models of calcium dynamics based on IP3 -sensitive calcium channel kinetics. Biophys J 1996; 70(1):246–63. [8] Keener, J. and Sneyd, J. (2008) Mathematical Physiology, Springer-Verlag, New York second edition. [9] Ressmeyer, A.-R., Bai, Y., Delmotte, P. F., Uy, K. F., Thistlethwaite, P., Fraire, A., Sato, O., Ikebe, M., and Sanderson, M. J. Human airway contraction and formoterol-induced relaxation is determined by Ca2+ oscillations and Ca2+ sensitivity. Am J Respir Cell Mol Biol 2010; 43(2):179– 91.


Calcineurin upregulates local Ca(2+) signaling through ryanodine receptor-1 in airway smooth muscle cells.

Receptor-activated calcium influx in human airway smooth muscle cells.

Cigarette Smoke and Estrogen Signaling in Human Airway Smooth Muscle.

Abnormal ryanodine receptor channels in malignant hyperthermia.

Calcium Sensitization Mechanisms in Gastrointestinal Smooth Muscles.

A deterministic model predicts the properties of stochastic calcium oscillations in airway smooth muscle cells.

Adrenoceptors in airway smooth muscle.

Defective Resensitization in Human Airway Smooth Muscle Cells Evokes β-Adrenergic Receptor Dysfunction in Severe Asthma.

Ovalbumin sensitization of guinea pig at birth prevents the ontogenetic decrease in airway smooth muscle responsiveness.

Ragweed sensitization-induced increase of myosin light chain kinase content in canine airway smooth muscle.

Actions of cAMP on calcium sensitization in human detrusor smooth muscle contraction.

Calcium antagonists and vascular smooth muscle cells in atherogenesis.

Relation between muscarinic receptor cationic current and internal calcium in guinea-pig jejunal smooth muscle cells.

Endothelin-induced calcium responses in human vascular smooth muscle cells.

Calcium and potassium currents in canine gastric smooth muscle cells.

Coupling mechanisms in airway smooth muscle.

Targeted transgenesis identifies Gαs as the bottleneck in β2-adrenergic receptor cell signaling and physiological function in airway smooth muscle.

Phosphoinositide metabolism in airway smooth muscle.

Involvement of Ca2+ Signaling in the Synergistic Effects between Muscarinic Receptor Antagonists and β₂-Adrenoceptor Agonists in Airway Smooth Muscle.

Airway smooth muscle and asthma.

Airway epithelial cells regulate membrane potential, neurotransmission and muscle tone of the dog airway smooth muscle.

Sex affects bone morphogenetic protein type II receptor signaling in pulmonary artery smooth muscle cells.

TGFβ Signaling-mediated MicroRNA Regulation in Vascular Smooth Muscle Cells.

TGF-β1 feedback circuit regulates airway remodeling in airway smooth muscle cells.