The methods used for the studies in this manuscript were adapted from (Hashad et al., 2017; Silpanisong et al., 2017; Thorpe et al., 2017; Reid et al., 2021) and are based on experimental animal, tissue preparation, confocal imaging, and electrophysiological studies as outlined in (Giang et al., 2016; Hashad et al., 2017; Shen et al., 2018; Hu et al., 2019; 2020; Leslie et al., 2021; Reid et al., 2021). In brief, studies were conducted in accordance with the Animal Welfare Act, the National Institutes of Health Guide for the Care and Use of Laboratory Animals ( https://grants.nih.gov/grants/olaw/Guide-for-the-Care-and-use-of-laboratory-animals.pdf), “The Guiding Principles in the Care and Use of Animals” approved by the Council of the American Physiological Society and were pre-approved by the Institutional Animal Care and Use Committee of Loma Linda University (IACUC-LLU).

As per our previous studies (Tao et al., 2015; Thorpe et al., 2017; Reid et al., 2021) pregnant ewes of a mixed Western breed were divided into low-altitude (normoxic) and high-altitude long-term hypoxic (LTH) groups. All ewes were obtained from Nebeker Ranch in Lancaster, CA at an elevation near sea level, 720 m. Normoxic control pregnant ewes were maintained near sea level at 720 m for the duration of their gestation. Animals for the LTH groups were held at Nebeker Ranch under normoxic conditions until 30 days gestation at which time the pregnant ewes were transported to the Barcroft Laboratory, White Mountain Research Station in Bishop, CA at an elevation of 3,801 m. Previous work shows that residing at the Barcroft laboratory results in a maternal arterial PO₂ of 60 ± 3 Torr and a fetal arterial PO₂ of 19 ± 2 Torr (Kamitomo et al., 1993). Animals remained at elevation for the remaining ∼110 days of gestation for pregnant ewes. Following this acclimatization period, ewes were transported (∼6 h drive) to Loma Linda University (LLU) for study at an elevation of 346 m. Once at LLU, LTH ewes were surgically instrumented with arterial and tracheal catheters. Based on frequent arterial blood gas sampling and adjustment of the rate of N₂ flow through the tracheal catheter, arterial PO₂ levels in adult sheep were maintained at ∼60 Torr for 2–4 days, mimicking the high altitude conditions until the day of study (Kamitomo et al., 1993). On the day of experimental study, ewes were induced with thiopental sodium (10 mg/kg iv), and then intubated. Anesthesia of the ewe was maintained via inhalation of 2%–3% isoflurane in O₂ for the duration of the surgery. Fetuses were delivered via hysterectomy. Generally, the male to female birth ratio is ∼1:1, however for the current studies only 2 female fetuses were identified in each group, precluding any substantive gender based analyses. Following delivery, fetal sheep were euthanized with an overdose of Euthasol (pentobarbital sodium, 100 mg/kg) and phenytoin sodium (10 mg/kg). Middle cerebral arteries were isolated from a total of 6 normoxic and 6 hypoxic fetuses.

Near term fetal brains of mixed sex were removed and placed in iced Balanced Salt Solution (BSS) of the following composition (mM): 126 NaCl; 5 KCl; 10 HEPES; 1 MgCl₂; 2 CaCl₂; 10 glucose; pH 7.4 (adjusted with NaOH). Middle cerebral arteries then were quickly dissected under normoxic conditions in BSS and stored in ice-cold BSS until used for experimental study.

Intracellular Ca²⁺ of middle cerebral arterial myocytes was measured in situ with a Ca²⁺ sensitive fluorescent dye (Fluo-4 AM, Cat No F14201, Invitrogen, Carlsbad, CA) using a Zeiss 710 NLO laser scanning confocal imaging workstation (Thornwood, NY) with an inverted microscope (Zeiss Axio Observer), using procedures based on previous studies (Harraz et al., 2014; 2015; Hashad et al., 2017; Reid et al., 2021). Fluo-4 AM was dissolved in DMSO creating a 1 mM stock solution. Arteries were placed in BSS containing 10 µM Fluo-4 with 0.1% Pluronic F-127 Cat No P6867 (Invitrogen) from a 20% w/v stock solution in DMSO for 1 h in the dark at room temperature (∼22–24 C). These arterial segments were subsequently washed with BSS for 30 min to facilitate dye esterification and were then cut into linear strips for experimental studies. Arterial myocytes were depolarized by treating arteries with a solution where there was an equimolar replacement of 30 mM NaCl for 30 mM KCl (30K) as we have performed previously (Hadley et al., 2012; Reid et al., 2021). The resting membrane potential based on examinations of middle cerebral arterial myocytes from cats, dogs, rats, and rabbits is predicted to range from roughly −70 mV to −40 mV (Harder, 1980; Fujiwara et al., 1982; Knot and Nelson, 1995; McPherson and Keily, 1995; Nelson et al., 1995; McNeish et al., 2010). Based on the Nernst reversal potential for potassium and a predicted intracellular potassium concentration of ∼ 140 mM (Wilson and Leblanc, 2005), the membrane is expected to depolarize to −40 mV; which we have previously shown to increase Ca²⁺ spark activity in basilar and pulmonary arterial myocytes (Reid et al., 2021; Hadley et al., 2012) and to elicit pulmonary arterial contractility (Papamatheakis et al., 2012). Arteries were then treated with 10 µM ryanodine in the presence of 30K for 30 min prior to imaging to reduce RyR activity (Janiak et al., 2001; Hadley et al., 2012).

Imaging was performed in arterial strips that were pinned to a Sylgard block in an en face orientation (Ellsworth Adhesives, Germantown, WI) with fine insect dissecting pins and placed into an open bath imaging chamber (Warner Instruments, Hamden, CT). The Fluo-4 was illuminated at 488 nm via a Krypton-argon laser. Emitted light was captured with a photomultiplier tube with a band limited range of 493–622 nm in both full frame and line scan imaging studies using a Zeiss Apochromat 63X Water Immersion 1.2 N.A. objective. All confocal imaging experiments were performed at Loma Linda University (346 m) under normobaric and normoxic conditions at room temperature (∼24°C) in the dark.

Time lapse recordings of multiple myocytes and subcellular Ca²⁺ recordings of individual myocytes were made in each arterial preparation as we have performed previously in tissues from multiple species (Hadley et al., 2012; Shen et al., 2018; Reid et al., 2021; Hashad et al., 2017; Hu et al., 2020). Whole-cell Ca²⁺ oscillatory activity of individual cells was assessed in time series recordings of 512 × 512 full frame images. These images were made with a lateral pixel size of 0.264 μm per pixel and an imaging depth adjusted to 5.4 μm, which is roughly the width of an arterial myocyte based on prior morphological studies in live and fixed cells (Hadley et al., 2012; Shen et al., 2018). These recordings ranged from 234 to 391 s at a frame rate of 1.28 Hz, which we have previously used for spatial and temporal examination of spontaneous whole-cell Ca²⁺ oscillations and stimulated responses (Shen et al., 2018; Reid et al., 2021; Leslie et al., 2021). Time series recordings were followed by 30–50 line scan recordings made along the edge of the sarcolemma at the axial center of the cell, which were made to measure subcellular Ca²⁺ spark activity. Each line scan was 18.9 s at a frequency of 529 Hz, with one line being made in each cell. The lateral pixel size for the 1,421 line-scan recordings that were made for these studies was 0.05 ± 0.03 μm per pixel. The pinhole was adjusted to an imaging depth of 2.5 μm, which is about 50% of the width of a myocyte based on prior morphological studies in live cells (Hadley et al., 2012; Shen et al., 2018).

Smooth muscle cells were enzymatically dissociated from fetal lamb middle cerebral arteries. Briefly, arteries were cut into ring segments and incubated (37°C, 10 min) in low-Ca²⁺ HEPES-buffered physiological salt (PSS) solution containing (in mmol/L) 140.0 NaCl, 5.0 KCl, 0.1 CaCl₂, 1.2 MgCl₂, 10.0 HEPES, and 10.0 glucose (pH 7.4). Vessels were then exposed to a 2-step digestion process that involved: (1) a 45 min incubation in low-Ca²⁺ HEPES-buffered PSS (37°C) containing 1.5 mg/mL papain (Worthington Biochemical; Lakewood, NJ), 1.5 mg/mL dithiothreitol (Millipore Sigma, Burlington, MA), and 1.5 mg/mL BSA (Millipore Sigma) and (2) a 45 min incubation in low-Ca²⁺ HEPES-buffered PSS (37°C) containing 1.5 mg/mL collagenase IV (Worthington Biochemical), and 1.5 mg/mL BSA (Millipore Sigma). After the enzyme treatment, tissues were washed with low-Ca²⁺ HEPES-buffered PSS. Single smooth muscle cells were released by gently inverting the tube(s) containing low-Ca²⁺ HEPES-buffered PSS and digested tissues several times. The cells were kept at 4°C and experiments were conducted within 6 h of cell isolation.

Spontaneous transient outward currents (STOCs) were recorded in the whole-cell configuration of the perforated patch-clamp technique using an EPC 10 patch-clamp amplifier with Patchmaster software (HEKA, Lambrecht/Pfalz, Germany) at room temperature as previously described (Hu et al., 2019). Briefly, cell suspension drops were placed in a recording chamber, and adherent cells were continuously superfused with HEPES-buffered PSS containing (in mmol/L) 140.0 NaCl, 5.0 KCl, 1.8 CaCl₂, 1.2 MgCl₂, 10.0 HEPES, and 10.0 glucose (pH 7.4). Only relaxed and spindle-shaped myocytes were used for recording. Micropipettes were pulled from borosilicate glass and had resistances of ∼ 5 megaohm (mΩ) when filled with a pipette solution containing (in mmol/L) 140.0 KCl, 1.0 MgCl₂, 5.0 Na₂ATP, 5.0 EGTA, and 10.0 HEPES (pH 7.2) with 250 μg/mL amphotericin B. Membrane currents were recorded while the cells were held at steady membrane potentials between −50 and 10 mV, which were increased in 10 mV increments of 1 min duration. STOCs were analyzed with Mini Analysis Program (Synaptosoft, Leonia, NJ) with a threshold for detection set at 10 pA. The currents were normalized to cell capacitance and expressed as picoampere per picofarad (pA/pF). Inhibitory effects of iberiotoxin and tetraethylammonium were examined in fetal hypoxic middle cerebral arterial myocytes at a holding potential of +10 mV. Because of inherent low STOC activity in myocytes isolated from normoxic animals these studies were only performed in hypoxic animals.

Ca²⁺ sparks recorded in cerebral arterial myocytes were analyzed using an automated approach with Sparklab 5.8 as recently described (Boskind et al., 2023). The analysis identified Ca²⁺ spark events and characterized Ca²⁺ spark morphology and spatial-temporal features of each event. Preliminary analysis worked towards the goal of optimizing the threshold used for Ca²⁺ spark detection and generating critical cutoffs for automated analysis. An evaluation of the threshold for Ca²⁺ spark responses above background noise was performed using various thresholds for spark detection including 0.5, 0.6, 0.7, 1.0, and 1.5 above background noise. A total of 214 true positive Ca²⁺ spark events were manually observed in 19 data traces imaged in myocytes of arterial segments isolated from 4 hypoxic fetuses, and comparisons were made to automated detection at various thresholds. From these data, the number of true positives, false positives, and false negatives were determined. This allowed for the calculation of the positive predictive value (PPV), sensitivity, and false discovery rate (FDR). The resultant analysis of Ca²⁺ spark events was made at a threshold of 0.7 as it provided a balance between the sensitivity of spark detection and the positive predictive value, which is the same cutoff we used previously (Boskind et al., 2023). Analysis was performed as we have done previously (Reid et al., 2021; Shen et al., 2018; Hu et al., 2019), and included the number of cells present with Ca²⁺ sparks, the frequency of activation as well as the event amplitude, full width at half maximum (FWHM), full duration at half maximum (FDHM), and the time constant for decay (Tau). These automatically detected features of the events were then filtered using an interquartile range (IQR) fencing method for amplitude (Lower and Upper), FWHM (Lower and upper), FDHM (Lower and Upper), and Tau (Lower and Upper), and statistical comparisons made (Boskind et al., 2023). The outlier fencing methods were based on the first and third quartiles (Q1, Q3) and the IQR. The lower inner fence was Q1 – (n × IQR), while the upper outer fence was Q3 + (n × IQR), where n represents 1.5. Percentiles, quartiles, and interquartile ranges were obtained via Prism 10.2.2. The total number of events examined along with the numbers examined after filtration with statistical fencing are provided in Table 1.

As described in some detail, regions of interest with Ca²⁺ oscillations were detected automatically post hoc using the LCPro plug-in for ImageJ (Shen et al., 2018; Reid et al., 2021). For presentation purposes, the fractional fluorescence intensity was automatically calculated using LCPro (Shen et al., 2018) following image registration and image cropping to a final size of 490 × 490 pixels. Regions of interest were detected for changes in fractional fluorescence in 1.56 μm (six pixel) defined regions, with the region size determined empirically as it allows for examination of fluorescence changes in individual cells as opposed to larger regions, which often overlap adjacent cells (Shen et al., 2018; Reid et al., 2021). For the purposes of the current study, erroneous data was removed using analytical techniques recently described (Boskind et al., 2023). This was achieved by using an outlier fencing approach to create upper and lower critical cutoffs along with an empirical approach of removing events shorter than 6 s or longer than 40 s. Further statistical analysis was performed, as described below. The total number of events examined along with the numbers examined after filtration with statistical fencing are provided in Table 2.

The spatial and temporal characteristics to the whole-cell Ca²⁺ oscillatory events were analyzed using customized processes outlined previously (Shen et al., 2018; Reid et al., 2021). This was performed to provide a better understanding of Ca²⁺ signaling networks within and between myocytes in the arterial wall. In brief, a cross correlation analysis for each Ca²⁺ event was performed in time and space for each region of interest detected by LC Pro (Shen et al., 2018). As per our previous studies, a correlation coefficient of r = 0.8 was used for all data sets (Shen et al., 2018; Reid et al., 2021). “Friends” were determined to be calcium transients that had a correlation coefficient greater than r = 0.8. “Neighbors” were defined as calcium transients occurring in myocytes within a radius of 100 pixels (∼26 µm), which was chosen because most all adjacent cells were within this distance.

Most reagents and chemicals were purchased from Sigma-Aldrich (St. Louis, MO) and Millipore Sigma (Burlington, MA). Fluo-4 AM and Pluronic F127 were purchased from Invitrogen, a division of Thermo Fisher (Waltham, MA) while iberiotoxin and ryanodine were purchased from Tocris (Ellisville, MO).

All time-series recordings were graphed, and statistical analyses performed with Prism 10.2.2. Contingency analysis was used in Figure 3A to examine Ca²⁺ spark activity based on the number of cells with and without events, with statistical differences examined by a chi square test. The data are displayed as the percentage of cells firing with Ca²⁺ sparks for each treatment and group. Histograms shown in Figures 4, 10 are the number of responding events in each bin and were tallied across all experimental treatments and groups. The large number of data replicates are suggestive that the data are derived from a population events. Summarized data for the Ca²⁺ spark events and Ca²⁺ oscillatory signals are presented as individual data points with mean and 95% confidence intervals. Data were evaluated for normality using a D’Agostino-Pearson test prior to any comparative statistical analysis. The specific test used is denoted in the figure legend or table. When the data were normally distributed a two-way repeated measures ANOVA with a Bonferroni’s multiple comparison test was used. When the data were not normally distributed then a Kruskal–Wallis one-way ANOVA with Dunn’s multiple comparison test. P values for ANOVA were adjusted for multiplicity for each comparison. P < 0.05 was considered statistically significant, unless otherwise noted. The P values are provided on the figure and denoted on the figure legend. The number of studies performed are provided as are the number of events, arteries, animals, or scans examined. The exact details are provided along with each figure and in Tables 1, 2.

The first series of studies were performed to examine Ca²⁺ spark activity in arterial myocytes. Preliminary analyses involved a determination of the optimal threshold for analysis of our dataset through a comparison of the spark detection capabilities of Sparklab 5.8 at multiple thresholds, using approaches based on our recent work (Boskind et al., 2023). Figure 2A showcases this comparative process in a representative line-scan image overlaid with markings denoting the individual Ca²⁺ events automatically detected at various thresholds by SparkLab and those that were manually detected by a trained observer. The image demonstrates how Ca²⁺ sparks with lower fluorescence intensities are increasingly indistinguishable from surrounding background noise at higher fluorescence thresholds, resulting in a decrease in the sensitivity of event detection. In comparison, at lower thresholds there was an elevation in the false positive rate as slight increases in fluorescence are more likely to be incorrectly recognized as a Ca²⁺ spark by SparkLab. To quantify the effectiveness and accuracy of each threshold to detect Ca²⁺ sparks, the sensitivity, positive predictive value (PPV), and false discovery rate (FDR) were calculated and graphed in Figure 2B. Notably, even at the least rigorous threshold of 0.5, the program detected Ca²⁺ sparks with a PPV and FDR of 80% and 19%, respectively. The FDR decreased to 12% and 7% percent at 0.6 and 0.7 threshold, respectively. Sensitivity incrementally decreased with increasing threshold through the removal of lower-quality events that were not substantially out of the noise floor. Based on this threshold analysis, we empirically proceeded with the analysis of our dataset at the 0.7 threshold as this criterion allowed for a balance between detecting “true” Ca²⁺ sparks while limiting the number of false positives. As we have recently shown, modifying the threshold has minimal or no impact on the spatial and temporal aspects to Ca²⁺ sparks (Boskind et al., 2023).

Delving into the central aim of our study, we explored the influence of long-term hypoxia on Ca²⁺ spark events in fetal sheep middle cerebral arterial (MCA) myocytes. Even under control conditions Ca²⁺ sparks occurred in 84% of the normoxic and 77% of the hypoxic myocytes (Figure 3A). Membrane depolarization with 30K increased the proportion of cells with Ca²⁺ sparks (Figure 3A) and their frequency of activation (Figure 3B), while the ryanodine treatment decreased the number of cells with Ca²⁺ sparks and their frequency in both normoxic and hypoxic groups, illustrating that these are classic Ca²⁺ sparks due to activation of ryanodine receptors (Jaggar et al., 2000; Janiak et al., 2001; Hadley et al., 2012). Long-term hypoxia, however, reduced the proportion of myocytes with Ca²⁺ spark activity relative to normoxia in control and 30 mM K⁺ (30K) groups while the frequency of Ca²⁺ spark activity was attenuated by long-term hypoxia in the 30K group.

The influence of long-term hypoxia on the spatial and temporal aspects to the Ca²⁺ sparks were then examined. Figure 4 shows histogram plots for the various morphological spark parameters that were examined including the amplitude, full-width at half maximum (FWHM), full-duration at half maximum (FDHM), and tau, where events were compiled across all groups and experimental conditions. Employing outlier fencing techniques, as outlined in our recent work (Boskind et al., 2023), we filtered the dataset to include only those events that met specific criteria. This involved applying a threshold based on the interquartile range (IQR) for data across all groups and experimental conditions multiplied by IQR 1.5 for the data depicted in each of the panels.

Once outliers were removed the various morphological parameters were measured and compared by experimental condition and for the effect of long-term hypoxia. As shown in Figure 5, Ca²⁺ spark events in myocytes of fetal myocytes were modified by the treatment and by long-term hypoxia. Figure 5A illustrates that in myocytes from normoxic animals, ryanodine treatment in the presence of 30K reduced Ca²⁺ spark amplitude. In myocytes from hypoxic animals, 30K with or without ryanodine increased Ca²⁺ spark amplitude. Relative to Ca²⁺ sparks from normoxic myocytes, 30K treatment also increased Ca²⁺ spark amplitude in myocytes from hypoxic animals. Figure 5B shows that Ca²⁺ spark width was less impacted by the treatments and by LTH. Ryanodine in the presence of 30K caused Ca²⁺ sparks to be narrower in myocytes from normoxic animals. Similarly, following 30K treatment Ca²⁺ sparks were narrower in myocytes from hypoxic relative to normoxic animals. Figure 5C shows that the duration of Ca²⁺ sparks was not systematically affected by the treatments or exposure to hypoxia. 30K preferentially decreased the duration of Ca²⁺ sparks in hypoxic animals while ryanodine in the presence of 30K shortened Ca²⁺ spark duration in normoxic myocytes. Under control conditions and following treatment with 30K and ryanodine, Ca²⁺ sparks in myocytes from LTH exposed fetuses had longer durations as compared to myocytes from normoxic animals. Still, Figure 5D shows that the tau for the decay of the Ca²⁺ signal was not affected in the same manner as the duration. The only difference in the time constant for decay of the Ca²⁺ signal was that the tau was increased in the 30K condition relative to control in myocytes from hypoxic animals.

We next examined the spontaneous activity of BK_Ca channels using patch-voltage clamp because Ca²⁺ sparks drive BK_Ca activity (Harraz et al., 2015; Hashad et al., 2018) and this axis is critical to feedback regulation of vascular tone (Jaggar et al., 2000; Ledoux et al., 2006; Jackson, 2021). Spontaneous outward currents are shown for middle cerebral arterial myocytes of fetal normoxic (Figure 6A) and hypoxic (Figure 6B) sheep for membrane voltages of −50 mV to +10 mV. As expected, the frequency of STOC activity increased with membrane depolarization in myocytes from both normoxic and hypoxic animals. The frequency of activity, however, was increased at more negative membrane potentials in the myocyte from the LTH animal. Figure 6C shows that the frequency of STOC activity was significantly increased by LTH. In myocytes from LTH sheep, the voltage dependence for STOC activation was shifted leftward along the voltage axis, such that there was greater channel activity at more negative membrane potentials. Figure 6D shows that STOC amplitude was also significantly greater at multiple membrane voltages following LTH.

Pharmacological studies were then performed to assess whether STOC activity was due to the activation of BK_Ca channels (Hu et al., 2020). Figure 7 shows that STOC frequency and amplitude were both reduced by either 100 nM iberiotoxin (Figures 7A–C) or 1 mM Tetraethylammonium (TEA) (Figures 7D–F).

Our investigation then aimed to assess the impact of membrane depolarization and long-term hypoxia (LTH) on Ca²⁺ oscillatory activity in middle cerebral arterial myocytes of fetal sheep. Figure 8A shows a maximum intensity projection of Fluo-4 fluorescence in myocytes of the middle cerebral arterial wall obtained from a time series recording. Figure 8B displays the Fluo-4 fluorescence over time recorded from two regions of interest (ROI’s) in individual myocytes, automatically detected and analyzed using LCPro, a custom analysis program (Francis et al., 2012; Shen et al., 2018). These recorded events, previously categorized as “medium duration,” were prevalent in whole-cell Ca²⁺ recordings, characterized by oscillations spreading uniformly through the myocytes and were empirically determined to last between 6 and 40 s (Reid et al., 2021). Shorter duration events appeared to be in discrete subcellular regions and indicative of Ca²⁺ spark events as previously described (Reid et al., 2021).

To better understand the holistic impact of LTH on Ca²⁺ oscillatory events we examined the number of events that occurred in sampled regions for each time series recording. These data are presented in Figure 9, where the total number of events in each video is presented. There was significant variability in the activity between the various recordings, which we believe limited our ability to determine if there are significant changes in cell activity due to hypoxia, membrane depolarization, or with ryanodine treatment. The only significant difference worth noting was that ryanodine in the presence of 30K reduced the number of events in normoxic MCA relative to membrane depolarization with 30K.

The investigation into Ca²⁺ oscillations encompassed an examination of the event amplitude, area under the curve (AUC), event duration, rise time, and decay. Figures 10A–E presents the graphical representation of events from distinct ROIs. We again performed outlier fencing techniques using IQR 1.5, for data across all groups and experimental conditions, with data outside the gray fences being removed from analysis. The duration of the oscillatory events was evaluated over the range of 6–40 s as opposed to using the IQR 1.5 values. This was done based on empirical knowledge of the duration of Ca²⁺ oscillatory signals, where we previously identified short duration events as Ca²⁺ sparks while long-duration Ca²⁺ events displayed uniquely distinct properties (Shen et al., 2018; Reid et al., 2021).

Quantitative analysis of the spatial and temporal aspects to the Ca²⁺ oscillatory events revealed nuanced effects of the treatments and experimental groups. The amplitude (Figure 11A) of the Ca²⁺ oscillations varied significantly across the treatment groups. The amplitude of the oscillations in the normoxic control group was significantly higher than in all other groups. Treatment with 30K decreased the amplitude of the oscillations as compared to controls. Interestingly, ryanodine treatment in the presence of 30K resulted in a further reduction in oscillatory amplitude. The area under the Ca²⁺ oscillation curve (Figure 11B) was modified by long-term hypoxia and by the treatment condition. Oscillations in the normoxic control group had the largest AUC, with oscillations in the other groups having significantly smaller areas. Rise time (Figure 11C) was longer in the control groups relative to 30K. The rise time for Ca²⁺ oscillations in the presence of 30K with ryanodine was shorter than control in the normoxic but not hypoxic condition. Oscillatory durations (Figure 11D) showed a pattern similar to the AUC, with events in the 30K group having a shorter duration compared to that of controls. The duration following ryanodine treatment was similar to that of control conditions. The decay in the oscillation (Figure 11E) for the controls was longer in the hypoxic as compared to normoxic group. 30K treatment shortened the decay time in both groups, while decay times increased following ryanodine treatment, being similar in length to control conditions.

Spatial and temporal correlations between the ROI Ca²⁺ transients were then examined. Increases in the correlation coefficient for temporally aligned Ca²⁺ oscillations denoted greater levels of event “friendship.” Temporally related correlations between the Ca²⁺ events were then identified by those events that had a correlation coefficient of r ≥ 0.8 (Figure 12A) (Shen et al., 2018). Results from a representative recording are provided in Figure 12. Figure 12A shows a maximum intensity projection from a time lapse recording with overlaid regions of interest that have varied levels of correlations. Figure 12B shows the fluorescence over time (black line) for the center ROI along with the fluorescence traces for those regions of interest that were temporally correlated. The spatial relationship between events was then evaluated (Figure 12B). ROIs that were within a 100 pixel (∼26 µm) radius of a reference ROI were denoted “neighbors,” as delineated by the larger white circle on the image in Figure 12B (Shen et al., 2018; Reid et al., 2021).

LTH and membrane depolarization had diverse effects on the numbers of friends, with summary data provided in Figure 13A. LTH decreased the number of friends under control conditions and following membrane depolarization. Membrane depolarization decreased the number of friends in the normoxic group but not the hypoxic group. Ryanodine treatment reduced the number of friends independent of altitude. The number of neighbors was also affected by LTH and membrane depolarization. Figure 13B shows that like friendship, the number of neighbors, which are those events within a 26 micron radius of a reference event, was reduced with LTH under control and 30K conditions. The number of neighbors was unchanged following membrane depolarization in the normoxic group but was increased in hypoxic conditions, while ryanodine reduced the number of neighbors in both groups. Figure 13C delineates the distance between friends. LTH reduced the distance between friends in both control and depolarized conditions. Membrane depolarization reduced the distance between friends in hypoxic but not normoxic conditions. Ryanodine treatment increased the distance between friends in hypoxic but not normoxic conditions.

Figure 13D provides the percentage of friends who are neighbors, illustrating the spatial and temporal relationships between Ca²⁺ oscillations. Overall, a majority of temporally correlated oscillatory events (friends) are closely associated (neighbors). Notably, these spatially correlated neighbor events that were also temporally correlated typically arose within the same cell as opposed to adjacent cells. Long-term hypoxia resulted in an increase in the percentage of friends that were neighbors in both the control and 30K conditions. Figure 13E presents the percentage of neighbors who are friends, which is dependent on the temporal relationship between Ca²⁺ oscillatory events that are spatially related. Events that occur in nearby regions of interest are not necessarily temporally correlated, though they are influenced by LTH and membrane depolarization. LTH increased the percentage of neighbors that are friends for control, 30K, and ryanodine conditions. Membrane depolarization reduced the percentage of neighbors who are friends in myocytes regardless of their exposure to LTH, while ryanodine increased the percentage.