Abnormal Ca2+ Spark/STOC Coupling in Cerebral Artery Smooth Muscle Cells of Obese Type 2 Diabetic Mice

Diabetes is a major risk factor for stroke. However, the molecular mechanisms involved in cerebral artery dysfunction found in the diabetic patients are not completely elucidated. In cerebral artery smooth muscle cells (CASMCs), spontaneous and local increases of intracellular Ca2+ due to the opening of ryanodine receptors (Ca2+ sparks) activate large conductance Ca2+-activated K+ (BK) channels that generate spontaneous transient outward currents (STOCs). STOCs have a key participation in the control of vascular myogenic tone and blood pressure. Our goal was to investigate whether alterations in Ca2+ spark and STOC activities, measured by confocal microscopy and patch-clamp technique, respectively, occur in isolated CASMCs of an experimental model of type-2 diabetes (db/db mouse). We found that mean Ca2+ spark amplitude, duration, size and rate-of-rise were significantly smaller in Fluo-3 loaded db/db compared to control CASMCs, with a subsequent decrease in the total amount of Ca2+ released through Ca2+ sparks in db/db CASMCs, though Ca2+ spark frequency remained. Interestingly, the frequency of large-amplitude Ca2+ sparks was also significantly reduced in db/db cells. In addition, the frequency and amplitude of STOCs were markedly reduced at all voltages tested (from −50 to 0 mV) in db/db CASMCs. The latter correlates with decreased BK channel β1/α subunit ratio found in db/db vascular tissues. Taken together, Ca2+ spark alterations lead to inappropriate BK channels activation in CASMCs of db/db mice and this condition is aggravated by the decrease in the BK β1 subunit/α subunit ratio which underlies the significant reduction of Ca2+ spark/STOC coupling in CASMCs of diabetic animals.


Introduction
More than 65% of patients with diabetes die from cardiovascular disease or stroke [1]. When considering age-adjusted incidence rates, type-2 diabetic patients are two-to five times as likely to suffer cerebral vascular disease or stroke compared with non-diabetic patients, a disparity that is seen in multiple racial/ geographic groups [2][3][4][5][6] and may result from abnormal cerebral artery tissue function. Interestingly, the incidence of stroke in type-2 diabetic patients is not associated with the duration of disease, smoking, fasting blood glucose, total cholesterol, lipoprotein concentrations, or hypertension [3,4,7]. Cerebral blood flow disturbances, impaired cerebral vascular reactivity, transient ischemic attacks and oxidative damage of cerebral vessels have been found in both type-2 diabetic patients [3] and experimental models [8][9][10] that could account for the higher incidence of diabetes-related stroke events [1][2][3][4][5][6][7]; however, the molecular mechanisms involved in cerebral artery dysfunction are not completely elucidated.
The db/db mouse, a genetic model of non-insulin dependent type-2 diabetes exhibits cerebral vascular dysfunction [9] that exacerbates brain damage, edema and inflammation after induced experimental stroke [11][12][13]. In addition to diabetes-related alterations found in cerebral vessels, vascular dysfunction is also present in mesenteric arteries [14][15][16], coronary arterioles [17], gracilis muscle arterioles [18,19], and aorta [20][21][22] of db/db mice. Cerebral arterioles of db/db mice show impaired response to vasodilators and reduced baseline arteriolar diameter [9]. Mesenteric arteries and gracilis muscle arterioles of db/db mice show impaired response to vasodilators, enhanced response to vasoconstrictors and enhanced basal myogenic tone [14][15][16]18,23]. Consistent with the observations of augmented vascular tone, the myogenic pressure-diameter of arteries harvested from db/db mice are smaller than the diameters of corresponding control arteries [14,18,19] and are not improved by the removal of endothelium [14,19]. Moreover, investigators have demonstrated impairment of endothelium-independent dilation in the presence of nitric oxide donors: in coronary arterioles and aorta of db/db mice in the presence of sodium nitroprusside [17,20], and in arteries of type 2 diabetic patients after administration of glycerin trinitrate [24,25]. All these data suggest that smooth muscledependent mechanisms are also responsible for the vascular dysfunction associated with type-2 diabetes. Furthermore, the disease appears to alter functional responses of resistance arteries not only at endothelial level but also in active smooth muscle layers.
In cerebral artery smooth muscle cells (CASMCs), spontaneous and local increases of intracellular Ca 2+ due to the opening of Ryanodine Receptors (RyRs), visualized as Ca 2+ sparks, activate large conductance Ca 2+ sensitive K + channels (BK channels) that generate spontaneous transient outward currents (STOCs) [26,27]. STOCs have a key role in the control of arterial tone by shifting the membrane potential towards less positive values, which in turn limits Ca 2+ influx through L-type Ca 2+ channels, diminishes global intracellular Ca 2+ concentration ([Ca 2+ ] i ), and opposes vasoconstriction [28][29][30]. Therefore, RyRs through Ca 2+ sparks and BK channels, by producing STOCs regulates arterial tone favoring vasorelaxation [29][30][31]. In addition, Ca 2+ spark generation is also regulated by Ca 2+ influx due to an indirect coupling between L-type Ca 2+ channels and RyRs [32]. The sarco/endoplasmic reticulum Ca 2+ ATPase (SERCA) participates in this indirect coupling by redirecting the Ca 2+ coming from the extracellular medium towards its luminal stores, located mainly inside the sarcoplasmic reticulum (SR) of CASMCs. Navedo and collaborators [33] demonstrated that L-type Ca 2+ currents and Ca 2+ sparklet activity -brief plasma membrane Ca 2+ fluxes due to the activation of clusters of L-type Ca 2+ channels-are enhanced in CASMCs of db/db mice; suggesting that the mechanisms that participate in the closure and tight regulation of L-type Ca 2+ channels could be impaired in diabetic CASMCs. Therefore, we sought to investigate whether BK channels and their functional pair, the RyRs, might contribute to worsening the condition of diabetic CASMCs.
Our data represent the first demonstration of altered Ca 2+ spark and STOC in CASMCs from db/db mice due to Ca 2+ spark alterations which lead to inappropriate BK channels activation, and the latter is aggravated by the decrease in the BK b1 subunit/ a subunit ratio also found in CASMCs of db/db mice.

Reduced Spatiotemporal Properties of Ca 2+ Spark in CASMCs of Diabetic Mice
Ca 2+ sparks of CASMCs represent the spontaneous and coordinated opening of an undefined number of RyRs within a cluster [26,30]. Their frequency and properties -i.e. amplitude, duration, size and mean rise rate-are used as markers of in situ RyR activity [34]. Since impaired function of RyRs might contribute to the vascular alterations in the cerebral arteries of diabetic db/db mice [8,9,33], our first aim was to study the in situ activity of vascular RyR in single, freshly isolated CASMCs of diabetic mice. Figure 1A shows representative confocal images of Ca 2+ sparks (right images) recorded in intact control and diabetic (db/db) CASMCs that conserve their typical relaxed spindle shape (left images). We found that the db/db CASMCs produced similar Ca 2+ spark frequency to control cells (in Hz: 0.860.1 vs. 0.860.1; or in events?s 21 ?mm 21 : 0.05360.008 vs. 0.04460.006, in 41 db/db cells and 43 control cells, respectively (Fig.1B), with no further modifications in the number of firing sites per cell, eager site probability and the maximum number of events within a same site ( Figure S1). The maintained Ca 2+ spark frequency in CASMCs contrasts with the decrease found in cardiac myocytes of the same mouse model [35]. However, db/db CASMCs presented a significant reduction in average Ca 2+ spark amplitude (Fig. 1C), full duration at half maximum (FDHM, Fig. 1D), full width at half maximum (FWHM, Fig. 1E Similar to previous observations in cardiac cells [35], histogram distributions of Ca 2+ spark parameters -plotted as a function of probability density (PDf)-were asymmetrical for both control and diabetic events (Fig 1F-H) and were fitted to a bimodal Gaussian function [35]. The amplitude distributions ( Fig 1F) showed 2 peaks: one around 1.25 F/F 0 and another of around 1.6 F/F 0 ; however, the population of events at larger amplitudes was drastically reduced in db/db compared to control CASMCs (43.7% in control cells to 5.9% in db/db cells, Table 1), which accounted for the decrease in the average Ca 2+ spark amplitude. In addition, the distribution of Ca 2+ spark durations showed 2 peaks that in db/ db cells were shifted towards shorter durations (in ms: first peak from 41.80 to 34.11 and second peak from 86.16 to 62.37, Table 1) with modified proportions (in %: first peak from 69.3 to 43.9 and second peak from 30.0 to 49.6). On average, these effects decreased the population of longer Ca 2+ sparks in db/db CASMCs (Figure1G). In the case of size (FWHM) distribution ( Fig 1H) the population of events changed from one peak distribution (mean at 2.77 mm) in controls to two peaks distribution (in mm: first peak at 2.03, second peak at 4.69, Table 1). Taken together, these results show a shift of the population to weaker, shorter and smaller Ca 2+ sparks in db/db CASMCs. Since Ca 2+ spark amplitude, duration and mean rising rate are good indicators of local release flux [34], we can conclude that although the total occurrence of Ca 2+ sparks are similar between db/db and control CASMCs, the total Ca 2+ mass (calculated by amplitude*duration*width) released through Ca 2+ sparks was decreased in db/db cells (in F/F 0 ?ms?mm: 317.0618.2 in 423 db/db events vs. 472.1630.8 in 460 control events, P,0.001). This could be due to a decreased number of RyR opening simultaneously in a cluster. We estimated RyR expression by western blots in aorta homogenates and found a decrease in db/db mice ( Figure 2).

Sarcoplasmic Reticulum Ca 2+ Loading CASMCs from Diabetic Animals
Sarcoplasmic reticulum (SR) Ca 2+ load is a major determinant of Ca 2+ release regulating RyR opening and, by consequence, Ca 2+ spark properties and activity. It is plausible, therefore, that a reduced SR Ca 2+ load could contribute to the reduction in Ca 2+ spark properties described from CASMCs of diabetic mice. Thus, we measured global SR Ca 2+ content by rapid application of caffeine (10 mmol/L). Global [Ca 2+ ] i responses were measured in CASMCs from control and db/db mice using fura-2 based Ca 2+ microfluorometry. Examples of such traces are shown in Fig. 3A, left traces. No significant differences in the peak amplitude of the caffeine-evoked [Ca 2+ ] i transients were observed (in nmol/L: 858.6681.6, n = 28 control cells vs. 731.4699.5, n = 18 db/db cells) (Fig. 3B, left bars) thus ruling out a difference on the SR Ca 2+ load at rest between db/db and control CASMCs underlying the weaker Ca 2+ sparks. To observe the time dependence of luminal Ca 2+ replenishment [36], we allowed 5 min of recovery and applied caffeine again. In this case, the peak amplitude of the second [Ca 2+ ] i transient was significantly reduced in the CASMCs of diabetic animals (in nmol/L: 675.9678.2, n = 16 control cells vs. 443.1641.9, n = 13 db/db cells, P,0.05) (Fig. 3B)

Decreased STOC Activity in CASMCs from Diabetic Mice
In CASMCs, Ca 2+ sparks activate large conductance K + (BK) channels, which play a critical role in controlling vascular tone [26,30]. We aimed to investigate whether the smaller Ca 2+ sparks in diabetic CASMCs had an impact on BK channel activity by analyzing spontaneous transient outward currents (STOCs). Intact CASMCs were patch-clamped and progressively depolarized from a holding potential of 250 mV to 0 mV in 10 mV voltage increments (Fig. 4A). STOC frequency was significantly decreased in db/db CASMCs compared to control cells across a range of voltages from 240 mV to 0 mV (Fig. 4B). However, sigmoidal fitting of the data revealed that the voltage dependence of STOC frequency was unaffected (voltage of half maximum activation, V 50 : 220.160.9, and 217.962.9 mV, in 9 control and 8 db/db CASMCs, respectively), suggesting that voltage sensitivity of BK channel activation is unaltered in type 2 diabetes. Amplitude was also significantly reduced in db/db CASMCs as compared to control cells, with statistical difference from 230 mV to 0 mV. However, linear fitting of these data demonstrated that the voltage dependence of STOC amplitude was diminished 0.55 fold in db/db CASMCs (slope: 0.4060.01, vs. 0.2260.02 pA/mV; Fig. 4C). The diminished STOC amplitude in the diabetic CASMCs was not due to a change in cell size, since there was no significant difference in membrane capacitance (an indicator of cell surface) between diabetic and non-diabetic CASMCs (9.562.4 pF for db/ db n = 10 cells vs.  The Proportion of b1 to a Subunit Expression of BK Channels is Decreased in db/db CASMCs We next sought to determine whether the diminished frequency and amplitude of STOCs in the diabetic CASMCs might be related to a change in BK channel composition. Vascular BK channels are composed of four pore-forming a-subunits and four accessory b-subunits, b1 being the predominant isoform expressed in vascular smooth muscle [37]. The a-subunit is the pore forming subunit, and the b1-subunit enhances the sensitivity of BK channels to Ca 2+ -dependent activation [37]. Downregulation or genetic disruption of b1 subunit has been implicated in the pathogenesis of diabetic vascular dysfunction [38][39][40][41] and arterial hypertension [42]. We determined protein expression of BK channel aand b1-subunits in aorta homogenates of control and db/db mice. Anti-actin antibody was used to normalize the loading. Representative immunoblots (Fig. 6A) and bar graphs (Fig. 6B) show that BK b1/a subunit ratio is 4.2-fold decreased in the diabetic mice (normalized ratio to actin: 1.060.37 vs. 0.2460.05, in n = 7 controls and n = 7 db/db tissue homogenates, P,0.05); suggesting that composition stoichiometry of BK channels is modified by the diabetic condition; with decreased proportion of BK channel b1-subunits to a-subunits. Separate analysis of BK a and b1 subunit expression levels were slightly but not significantly modified (BKa normalized ratio to actin: 1.060.3 in controls vs.

Reduced STOC/spark Coupling in CASMCs of Diabetic Mice
It has been shown that Ca 2+ sparks have a vasorelaxing effect by activating STOCs, which hyperpolarizes the cell, leading to inactivation of VDCCs. The STOC/Ca 2+ spark coupling ratio was indirectly estimated by dividing the number of STOCs at near resting membrane potential (240 mV) by the number of spontaneous Ca 2+ sparks in single quiescent cells. Fig. 7A shows that this coupling ratio was decreased in db/db CASMCs, suggesting that many of the Ca 2+ sparks produced in the diabetic CASMCs lack STOCs, contrary to what is observed in normal vascular myocytes [29]. The estimated reduced coupling ratio may have a considerable impact on the deregulation of vascular tone under the diabetic condition.

Discussion
In this work we present the first compelling evidence of the RyR and BK channel functional loss in native CASMCs of db/db mice. We have demonstrated: 1) Ca 2+ spark properties are significantly decreased in the CASMCs of diabetic animals; 2) STOC frequency, amplitude, kinetics and area were further reduced; and 3) diminution of the BK channel b1/a subunit ratio, resulting in an abnormal Ca 2+ spark-STOC coupling in db/db CASMCs.
Taken together, our data suggests the scenario schematized in Fig. 7B where the triggering of STOCs is not tightly controlled by Ca 2+ sparks in db/db CASMCs, probably due to 1) smaller and shorter Ca 2+ sparks with reduced amplitude than then could be STOCless; 2) a compromised SR Ca 2+ load recovery; and 3) decreased STOC frequency, amplitude, area, time-to-rise and time-to-decay. The latter is probably due to reduced Ca 2+ sparks because of decreased RyRs expression and further enhanced by the altered subunit composition of BK channels. Taken together, the reduced efficacy of Ca 2+ sparks to trigger STOCs might contribute to an increase in vascular tone of db/db arteries.  Previous studies have shown that the function of cerebral arteries is impaired in experimental and genetic models of type-2 diabetes [8][9][10]41]. Specifically, the cerebral vascular dysfunction found in the db/db mice is characterized by reduced baseline arteriolar diameter, increased basal tone and reduced response to acetylcholine, in part due to enhanced Rho-kinase activity, superoxide production [9] and overall oxidative damage [8] that worsens brain damage, edema and inflammation after induced experimental stroke [11][12][13]. In the current study, we have analyzed intracellular Ca 2+ handling and STOCs which might contribute to these alterations, since it has been previously shown that spontaneous Ca 2+ sparks contribute to vascular relaxation by activating STOCs.
We found that the overall Ca 2+ spark frequency was not significantly altered, but the Ca 2+ spark amplitude, duration, width, and as a consequence Ca 2+ spark mass were decreased in db/db CASMCs. To our knowledge, this is the first study of Ca 2+ sparks in CASMCs of type 2 diabetes. In type 1 diabetes, Ca 2+ sparks frequency has been found to be unaltered [40] or increased [39], while their amplitudes were found to be increased [39,40]. Although both type 1 and type-2 diabetes are characterized by hyperglycemia, the mechanism of disease is quite different, which may explain the differing results in Ca 2+ spark amplitude.
It has been previously established that decreasing SR Ca 2+ load in vascular myocytes reduces Ca 2+ spark properties [43]. We did not find a significant decrease in the SR Ca 2+ load in the conditions of Ca 2+ sparks recording (Fig. 3), ruling out this possibility. However, the recovery of SR Ca 2+ load after 5 min of depletion was impaired in db/db mice, which could be indicative of an underlying defect in SERCA activity. In fact, it has been documented that cytokines (i.e. interleukin-1 band c-interferon), which are increased in the db/db mice, depress SERCA function [44]. On the other hand, reduced number of RyRs that simultaneously open in a cluster may account for the reduction in Ca 2+ spark amplitude, duration, spatial spread, and mass of the Ca 2+ spark. This could be supported by the reduced RyR expression in db/db arteries (Fig. 2).
Ca 2+ sparks in vascular smooth muscle cells contribute to relaxation by activating BK channels. Diabetic CASMCs show a significant decrease in STOC frequency, amplitude, time-to-rise, time-to-decay and area ( Fig. 4 and 5), as in type 1 diabetes [39,40]. STOCs are outward (hyperpolarizing) currents mediated by BK channels. The K + efflux brings the cell membrane to more negative potentials, limiting the activation of voltage-dependent Ca 2+ channels. This in turn decreases Ca 2+ influx, inducing relaxation [26][27][28][29][30][31]. Hence, the decrease in Ca 2+ sparks properties, STOCs and Ca 2+ sparks-STOCs coupling could underlie or contribute to depolarization under certain conditions in vivo.
Because STOCs are activated by Ca 2+ sparks, and the occurrence of Ca 2+ sparks is not significantly altered, the reduction in STOCs could be interpreted as a decreased ability of the Ca 2+ sparks to activate STOCs in db/db CASMCs (Fig. 7A). There are multiple potential explanations for this uncoupling. It is possible that the smaller Ca 2+ sparks are less effective in activating STOCs. Indeed, after careful analysis of Ca 2+ spark amplitude, duration, and width PDf (Fig. 1F-H), we found that the global decrease in Ca 2+ spark properties was related to a decrease in the population of larger Ca 2+ sparks. Additionally, the reduction in STOCs could be due to the relative decrease in b1 channel subunit (Fig. 6) [38][39][40][41]45], since it has been shown that the b-subunit plays a major functional role in defining both the Ca 2+ dependence of activation and availability of BK channels [37]. In this regard, BK channels of coronary arteries from Zucker diabetic fatty rats showed impaired Ca 2+ dependent activation attributable to a decrease in BK channel b1-subunit expression [38]. We found a decrease in the b1/a BK subunit proportion (Fig. 7), suggesting that in db/db mice, some BK channels may be devoid of b1-subunit. Additional biochemical studies of type-1 diabetic animals also support the evidence that BK channel b1-subunit relative to BK channel asubunit protein expression is reduced in arterial tissues [39,40]. Interestingly, similar findings have been reported in CASMCs of spontaneously hypertensive rats [42] underlying a common mechanism to decrease STOC/Ca 2+ spark functional coupling. Additionally, a key finding of this study is that the function of BK channels is impaired in CASMCs of type-2 diabetic mice, but that this is not attributable to an altered voltage dependence of BK channel activation. Although we cannot exclude the possibility that the dysfunction of BK channels found in diabetic CASMCs could be due to high glucose-mediated oxidative modulation [46], we suggest that Ca 2+ sparks with diminished amplitude, size, duration, and mass may be less efficient in activating STOCs. Indeed, the STOC/Ca 2+ spark coupling efficiency was significantly reduced in db/db cells (Fig. 7A), which suggests that spark/ STOC coupling plays a key role in contributing to the previously well documented vascular dysfunction associated with this model of type 2 diabetes [9]. Further, the decreased STOC activity in db/ db mice could underlie the increased activity of L-type Ca 2+ channels reported by Navedo et al. [33].
In summary, we have dissected local Ca 2+ signaling-STOC activity in CASMCs from db/db mice and their controls. We have shown that the alteration in RyR activity, manifested as a decrease in the larger Ca 2+ spark population, is involved in reduced STOC properties, with consequent dysfunction. Thus, defective crosstalk between RyRs and BK channels may be important contributors to the vascular pathology of diabetic individuals and, as such, are potentially attractive targets for therapeutic intervention.

Materials and Methods
All experiments were carried out according to European Union Council Directives (86/609/EEC) for the care of laboratory animals. Animal Protocol was approved by the Comité Régional d'Ethique sur l'expérimentation animale of Languedoc-Roussillon on the Use and Care of Animals (authorization B34-172-16 for animal facility manager).

Mice
The diabetic (db/db) mice used in the current study were on a C57BL/KsJ genetic background. We used male animals at the age of 14 to 16 weeks. db/db mice were significantly heavier than WT

Ca 2+ spark Recordings
Spontaneous and local Ca 2+ release events were recorded as previously reported [47]. Fluo-3 loaded CASMCs (50 ml of cell suspension) were allowed to adhere to the bottom of a glass coverslip in a perfusion chamber. The cells were perfused with physiological saline solution (PSS) at room temperature before starting the experiments. Ca 2+ sparks were recorded at room temperature with a laser scanning confocal microscope (Zeiss, LSM510 META) equipped with an x63 water-immersion objective (N.A. 1.2) in the line scan mode (5 images of 1000 lines at speed of 1.92 ms/line). Fluo-3 was excited at 488 nm with an Argon laser (3% intensity), and emission measured at above 510 nm. Ca 2+ sparks were reconstructed by stacking consecutive line scans and performing a time-intensity plot. Ca 2+ sparks were automatically detected and properties of amplitude (F/F 0 ), duration (FDHM, Full duration at half maximum, in ms) and width (FWHM, Full width at half maximum, in mm) measured with a custom-made program running in IDL 5.5 software (Research Systems Inc.), with a detection threshold of 4.36S.D [36]. Images of Ca 2+ sparks were normalized by dividing the fluorescence intensity of each pixel (F) by the average resting fluorescence intensity (F 0 ) to generate an F/F 0 image. Mean rate of rise was determined as reported by Shen et al [34].

[Ca 2+ ] i Transients in CASMCs
Myocytes were loaded with 2.5 mmol/L fura 2-AM. Cells were transferred to a bath perfusion chamber and continuously superfused with PSS. Two caffeine-induced [Ca 2+ ] i transients were elicited consecutively with a 5-minute interval in between to allow refilling of intracellular Ca 2+ stores. Images of Fura-2 fluorescence were captured by a cooled CCD camera (Photometrics, USA) with an oil-immersion x40 objective mounted on an inverted microscope (Axiovert, Zeiss, Germany), and acquired at 510 nm emission after excitations at 340 nm and 380 nm using a lambda-DG4 excitation system (Sutter Instrument Company, USA). Ca 2+ signals were analyzed using Metafluor software (Universal Imaging Corporation, USA). [Ca 2+ ] i (in nmol/L) was calculated from the ratio of 340/380 nm after correcting for background. Fluorescence emission ratios were converted to [Ca 2+ ] i according to the Grynkiewicz equation [48], using a dissociation constant for fura-2 of 239 nM. Values of 0.32, 5.7, and 10.3 obtained from fura-2 in situ calibration were used for Rmin, Rmax, and b, respectively.

Electrophysiological Recordings
STOCs were recorded in the whole cell configuration of the patch clamp technique using solutions and protocol previously described [48]. The pipette solution (in mmol/L): 80 Kglutamate, 5 NaCl, 40 KCl, 1 MgCl 2 , 3 MgATP, 0.1 NaGTP, 0.05 KEGTA, 20 HEPES, pH 7.2 with KOH. CASMCs were superfused with PSS and clamped at 250 mV or stepped from 250 to 210 mV in 10 mV increments. Membrane capacitance (Cm) was determined from the current amplitude elicited in response to a hyperpolarizing voltage pulse from a holding potential of 280 mV (duration, 10 ms; amplitude, 10 mV). Currents were filtered at 500 Hz and digitalized at 2 kHz (500 ms/point). STOC analysis was performed off line, using the event detection tool of Clampfit 9.2 (Axon Instruments, Inc); where both STOC time-to-rise and time-todecay were measured at half peak of STOC in ms.

STOCs and Ca 2+ Spark Analyses
Data are expressed as the mean 6 SE. Statistical significance was assessed using SigmaStat 3.0 software by Student's t-tests, one way ANOVA test or by nonparametric Mann-Whitney Rank-sum test for cases in which data failed the Shapiro-Wilk normality test. Values of P,0.05 were considered statistically significant. The probability density function analysisPDf a ð Þ~N a ð Þ Ntotal : D , where N(a) is the histogram distribution of Ca 2+ spark parameter a, N total is the total number of Ca 2+ sparks, and D is the bin width. Bimodal Gaussian distributions were done as previously reported [35]. Figure S1 A, Average of sites where Ca 2+ sparks were recorded within the same cell during the recording period (9.6 seconds).

Supporting Information
Firing sites were counted as the sites where we recorded at least one Ca 2+ spark. B, Ca 2+ spark frequency within each firing site reported as number of events recorded within each site/s. C, Probability in each cell to present sites that fire repetitively. D, Maximum number of Ca 2+ sparks recorded within the same site. N = 43 for control CASMCs (white bars) and n = 41 for db/db cells (gray bars). (DOCX)