A new method to detect rapid oxygen changes around cells: How quickly do calcium channels sense oxygen in cardiomyocytes?

Abstract

Acute hypoxia is thought to trigger protective responses that, in tissues like heart and carotid body, include rapid (5–10 s) suppression of Ca²⁺ and K⁺ channels. To gain insight into the mechanism for the suppression of the cardiac l-type Ca²⁺ channel, we measured O₂-dependent fluorescence in the immediate vicinity of voltage-clamped cardiac cells subjected to rapid exchange of solutions with different O₂ tensions. This was accomplished with an experimental chamber with a glass bottom that was used as a light guide for excitation of a thin ruthenium-based O₂-sensitive ORMOSIL coating. Fluorescence imaging showed that steady-state Po₂ was well controlled within the entire stream from an electromagnetically controlled solution “puffer” but that changes were slower at the periphery of the stream (τ_1/2 ∼ 500 ms) than immediately around the voltage-clamped myocyte (τ_1/2 ∼ 225 ms) where, in turn, firmly attached cells produced an additional local delay of 50–100 ms. Performing simultaneous voltage clamp and O₂ measurements, we found that acute hypoxia gradually and reversibly suppressed the Ca²⁺ channel (Ca_V1.2). Using Ba²⁺ as charge carrier, the suppression was significant after 1.5 s, reached ∼10% after 2.5 s, and was nearly completely reversible in 5 s. The described fluorescence measurements provide the means to check and fine tune solution puffers and suggest that changes in Po₂ can be accomplished within ∼200 ms. The rapid and reversible suppression of barium current under hypoxia is consistent with the notion that the cardiac Ca²⁺ channel is directly modulated by O₂.

biological organisms respond to environmental stress with regulatory processes that act on different timescales and serve to both counteract immediate injury and maintain homeostasis. Within the heart, few stressors are as important and detrimental as abnormal oxygen levels. Prolonged hypoxia, arising from conditions such as ischemic heart disease, may overwhelm cellular capacity to resist damage and result in pathophysiological function (4). Ischemic responses include acidification, changes in reactive oxygen species (ROS) levels, decreasing ATP concentration, and ultimately cell death (3). Hypoxia also occurs in numerous other conditions such as myocardial infarction, cardiac hypertrophy, and microvascular disease (10). Understanding the mechanisms by which myocardium senses and responds to hypoxia may provide insight into the design of novel drugs and therapeutic approaches.

While it has long been known that cardiac myocytes respond rapidly to hypoxia by suppressing l-type Ca²⁺ channel current (carried by Ca²⁺ or Ba²⁺), this response is not yet fully understood (8). We have previously suggested that the Ca²⁺ channel functions as an acute oxygen sensor for the heart, in which the decrease in calcium current (I_Ca) triggered by hypoxia involves the Ca²⁺/CaM binding domain of the channel's C-terminus (18, 23). Others have suggested that the suppressive effects of hypoxia on I_Ca is mediated by changes in production of mitochondrial ROS, which in turn interact with thiol groups on the C-terminus of the Ca²⁺ channel (12–14). In part, these alternative possibilities may be distinguished through the kinetics of the response of l-type Ca²⁺ channels to hypoxia.

To address this issue, we developed a novel method for sensing dissolved O₂ in the immediate vicinity of patch-clamped myocytes. Using a ruthenium-doped ORMOSIL coating as an O₂ sensor (5, 20), we prepared coverslips, the fluorescence of which was rapidly and reproducibly quenched by the binding of dissolved O₂. Plating cardiac cells onto these coverslips, we verified that hypoxia rapidly suppressed current carried by l-type Ca²⁺ channels and established that the suppressive effect of acute hypoxia results from hypoxic conditions lasting as little as 1.5 s in duration.

MATERIALS AND METHODS

Isolation of rat cardiac myocytes.

Cardiac myocytes from the right ventricle of male Sprague Dawley rats were isolated as previously described (17). Protocol for animal handling and cell isolation (AR 2791) was approved by the Institutional Animal Care and Use Committee of the Medical University of South Carolina according to national legislation and guidelines.

Hearts were surgically extracted from animals anaesthetized by isoflurane (3%, vaporized), retrogradely cannulated through the aorta, and washed for 5 min by perfusion of calcium-free Tyrode's solution (in millimoles: 136 NaCl, 30 taurine, 10 HEPES, 10 d-glucose, 5.4 potassium l-glutamate, and 1 MgCl₂ at pH 7.2 using NaOH) using a peristaltic pump at 7 mL/min. After beats stopped completely, hearts were enzymatically digested by a mixture of collagenase type 2 (1.32 mg/mL, 270 units/mg, Worthington) and protease type XIV (0.26 mg/mL, ≥ 3.5 units/mg, Sigma) for 15–20 min. Postdigestion, hearts were washed with 0.2 mM Ca²⁺ Tyrode's solution for 8 min. All solutions were maintained at 37°C using a heating circulator bath, and perfusion steps were conducted under a heating lamp to maintain heart temperature. The right ventricle was dissected from the digested heart and cells were released from the tissue by gentle manual dissociation using forceps. Cells were plated onto oxygen sensor-doped ORMOSIL-coated coverslips in 0.2 mM Ca²⁺ Tyrode's solution and were allowed to settle for a minimum of 45 min prior to patch clamp experiments.

Coverslip coating materials and procedure.

Tetramethylorthosilicate (TMOS, 98%), trimethoxy(propyl)silane (TMPS, 97%), ACS grade hydrochloric acid (HCl, 37%), nitric acid purified by redistillation (HNO₃, 70%), and absolute ethanol (EtOH, 200 proof) were purchased from Aldrich. Tris(4,7-diphenyl-1,10-phenanthroline)ruthenium(II) dichloride complex [Ru(dpp)₃]Cl₂, ≥95% (HPCE) was obtained from Fluka.

Method for the synthesis of [Ru(dpp)₃]Cl₂-doped ORMOSIL coating was adapted from those previously described (5, 20). Twenty four hours prior to ORMOSIL preparation, glass coverslips (22 × 40 mm, Corning, 2935–224) were placed into a bath of HNO₃ to prepare the surface for coating. Immediately prior to coating preparation, coverslips were removed from HNO₃ bath, washed by successive rinses with EtOH and distilled water, and air dried. OROMSIL matrix was prepared by mixing TMOS (100 μL, 0.672 mmol) and TMPS (160 μL, 0.908 mmol) together with a magnetic stir bar in a 25 mL round bottom flask. Upon pipetting both solutions into the reaction vessel, HCl (200 μL, 0.01 M) was immediately added in a dropwise manner. The solution was stirred for 1 h in a closed flask. Meanwhile, a 5 mM stock solution of [Ru(dpp)₃]Cl₂ was prepared by combining 1.00 mg (855 nmol) with 171 μL of EtOH, then vortexing until homogeneous. The 5 mM [Ru(dpp)₃]Cl₂ stock solution was combined with the completed ORMOSIL matrix in a microcentrifuge tube in a 1:1 ratio, then mixed by vortexing. Immediately, a portion of this mixture was pipetted (4 μL) onto a prepared glass coverslip and spread by drawing a second coverslip over its surface. Coverslips remained in ambient conditions for a minimum of 24 h before plating with cells.

Patch clamp recordings.

Freshly isolated cardiac myocytes plated onto [Ru(dpp)₃]Cl₂-doped ORMOSIL-coated coverslips were voltage-clamped in the whole cell configuration on an inverted microscope using a Dagan 8900 Patch Clamp amplifier (Dagan, Minneapolis, MN), Digidata 1440A digitizer and pClamp 10.2 software (Axon Instruments, Union City, CA). Ba²⁺ currents (I_Ba) through l-type Ca²⁺ channels were measured using patch pipettes with 2–5 MΩ resistance filled with an internal solution containing (in millimoles): 132 CsCl, 11 d-glucose, 11 HEPES, 11 TEACl, 5.5 MgATP, 5.5 NaCl, 2.2 EGTA, and 0.92 CaCl₂ at pH 7.2 using CsOH. Ba²⁺ was chosen as the charge carrier because, unlike I_Ca, the suppressive effect of hypoxia on I_Ba is not mitigated by phosphorylation (23). The extracellular Tyrode's solution contained (in millimoles) 137 NaCl, 10 HEPES, 10 d-glucose, 2 BaCl2, and 1 MgCl₂ at pH 7.4 using NaOH.

Solution surrounding the voltage-clamped cell was quickly exchanged using a digitally controlled five-barrel rapid perfusion system which, judged by electrophysiological responses, exchanged solutions within ∼50 ms (6). To simulate hypoxic conditions, a solution of extracellular Tyrode's was bubbled with 100% N₂, yielding a solution containing ∼1% O₂ measured using an oxygen-sensitive electrode. When not experiencing hypoxia, cells were bathed in Tyrode's solution bubbled with 100% O₂ or equilibrated with ambient air (∼20% O₂).

I_Ba through calcium channels was measured using a 30 ms membrane depolarizations from −50 to 0 mV pulsed every 5 s. Cells were exposed to hypoxic solution every third depolarization for a duration ranging from 2.5 to 0.5 s. Measured currents were filtered at 10 kHz. All experiments were done in room temperature.

Oxygen measurement.

The retention of [Ru(dpp)₃]Cl₂'s luminescent properties was verified by excitation (measured emission λ = 610 nm) and emission (excitation λ = 470 nm) spectra of sensor-coated coverslips at room temperature using BioTek Synergy™ H1 Hybrid Multi-Mode Microplate Reader (Fig. 1A). Oxygen-quenching fluorescent characteristics were confirmed by rapidly dispensing O₂ and N₂ saturated H₂O onto a sensor-doped ORMOSIL-coated six-well plate.

Fig. 1.

Fluorescence response of the ruthenium-based ORMOSIL coating to changes in O₂ tension. A: excitation and emission spectra of the O₂-sensing coating under solution perfused with hypoxic (N₂, red), normoxic (ambient, green, Po₂ ∼0.2 bar), and 100% O₂ (O₂, black, Po₂ ∼1 bar) solutions. B: diagram of experimental configuration. A rapid perfusion system ending in a thin polyethylene tube (0.28 mm inner diameter, Intramedic) was used to direct a stream of solution to an area of 800 μm in diameter that was imaged on a Leica microscope in the epi-fluorescence mode. C and D: distribution (C) and time course (D) of the O₂-dependent fluorescence from coated coverslips during rapid exchange of solutions with different O₂ tensions (a: normoxic; b: hypoxic; c: hyperoxic). C shows sample images (a, b, and c) with the steady-state fluorescence distributions produced by the three test solutions and also color-coded regions of interest (ROI) relative to the sketched outlet of the perfusion system. The colored traces in D show the kinetics of the fluorescence response measured in the corresponding similarly coded regions of interest.

During voltage clamp experiments, the ruthenium-based sensor was excited by evanescent illumination at 473 nm from a 400 mW diode laser. In this configuration the coated coverslips (22 × 40 mm, Corning, 2935–224) were used as an exchangeable bottom and light guide in the experimental chamber and light was admitted along the short edge via optical fibers (Light guide-based TIRFM, TIRF Labs, Cary, NC). Emitted light (>515 nm) from the paracellular vicinity was detected by a photomultiplier tube behind an aperture that limited the detection area to the voltage-clamped cells and its immediate surroundings (∼100 μm radius around cell) or wider area. Using this optical system, we verified that our perfusion system was capable of changing the O₂ concentration around the cell with a time to half maximum (τ_1/2) of 225 ms ± 14 ms (± SE), n = 6.

Fluorescence distribution was additionally verified in the epi-illumination mode with a Leica microscope (DMI6000 B, Buffalo Grove, IL) equipped with a iXon3 897 camera (Andor Technology, Belfast, UK). Fluorescence images [F(x,y,t)] were typically recorded at 20 Hz and were displayed ratiometrically [F(x,y,t)/F_normoxic (x,y)] (2) after division with the normoxic steady-state fluorescence distribution [F_normoxic(x,y)]. This distribution was either measured directly (Fig. 1C) or was calculated as a weighted average [F_normoxic ≅(2* Fn₂ + Fo₂)/3, Fig. 2] from the steady-state fluorescence distributions measured with solutions bubbled with N₂ (Fn₂) or O₂ (Fo₂). The weights (2:1) were derived from the average spectrofluorometric and imaging responses [(Fn₂ - F_normoxic)/(F_normoxic - Fo₂) = 2.4 ± 0.5, n = 6].

Fig. 2.

Local changes in O₂-dependent fluorescence immediately following solution exchange. The insets show a bright field image (BFI) of the puffer outlet with several cells in front of it, regions of interest (ROI), and two sequences of numbered sample frames measured at 20 Hz when switching from fully oxygenated to hypoxic solution (#146–154) and back (#183–191). Notice that the measured O₂ changes occur earlier at the center of the stream hitting the coverslip (red and green ROIs) than at the periphery and that the coating under the cells (red ROI, white arrows) responds with a delay of 50–100 ms compared with the immediately surrounding regions of the coverslip (green ROI). The sample frames were measured ratiometrically relative to the normoxic steady-state fluorescence distribution as described in the Methods section. The color code is identical to Fig. 1 (blue: 100% O₂; green: normoxia; yellow/orange: hypoxia). The lower row of images show the delayed spread of hypoxia (consecutive frames 1, 2, and 3) under an intact cardiomyocyte (BFI-2).

Statistical analysis (or data analysis).

Suppression of I_Ba was calculated by taking the difference between the peak current during hypoxia and the projected current with no hypoxia, accounting for rundown (Fig. 4). Student's t-test was performed to check for significance between the control and experimental groups for each duration of hypoxia (0.5, 1, 1.5, 2, and 2.5 s). Less than 5% probability of equality between groups is denoted by * and less than 1% is indicated by **.

Fig. 4.

Effect of brief, transient hypoxia on I_Ba. A: sample experimental and control traces showing the effect of 2.5 s of hypoxia on I_Ba. Control and experimental traces were recorded in the same cell; insets A and B represent data from different cells. B: relationship between the duration of hypoxia experienced by the cell and I_Ba. Current density decreases with time as a function of both hypoxia and rundown. Hollow points indicate depolarizations where the cell was in hypoxic conditions, duration indicated above point. Black points indicate depolarizations where the cell was being perfused with 100% O₂ solution. Black curve indicates proposed pattern of rundown. I_Ba suppression by hypoxia was quantified as the difference between proposed rundown curve and the suppressed current as indicated by a dashed line.

RESULTS

Oxygen measurement.

After preparing the Ru-based O₂-sensitive ORMOSIL coating, we examined the magnitude, uniformity, and time course of the responses of the coated coverslips to solutions with different O₂ tensions. As a first step, we coated the bottom of a polystyrene six-well plate and used a spectrofluorometer to measure the absorption and emission spectra, which displayed up to a threefold reduction in fluorescence intensity when the superfusing solution was changed from one that was bubbled with 100% N₂ to another bubbled with 100% O₂ (Fig. 1A). Solution equilibrated with ambient air produced intermediate responses. Next we reproduced the geometry of biological experiments where a small area of the coated coverslips was exposed directly to rapid flow of solutions containing different concentrations of O₂ (Fig. 1B). Figs. 1C and 1D show the steady-state images and the time course of the fluorescence signals measured when Po₂ of the superfusing solutions was changed between 100%, 0_% (100% N₂), and 20% (atmospheric) oxygen. The O₂-dependent fluorescence signals changed most rapidly directly in front of the outflow tube [green trace in Fig. 1D and region of interest (ROI) in Fig. 1C], spread with a slight delay in concentric circles (red and blue), and had a relatively well defined outer limit that in the shown sample frames (b and c in Fig. 1D) is seen only in lower left-hand corner (orange). Rapidly exchanged solutions, shown in Fig. 1D, quickly reach a steady-state level but consistently showed slight delays in response time in areas farther away from the central puffing region.

Figure 2 shows images of O₂-dependent fluorescence recorded at 20 Hz from an oxygen-sensing coverslip plated with cells as seen in the bright field image (BFI). Selected frames recorded at the onset (#146–154) and termination (#183–191) of a brief hypoxic interval confirmed that the solution directly in front of the puffer outlet changes first (red and green ROIs and curves). This change in fluorescence expanded radially (blue ROI) but terminated at the interface of superfusing stream and the bath solution (orange ROI). Analysis of the frames recorded immediately after the solution changes (frames #147, 148, 184, and 185) showed that there was a delay of ∼100 ms in O₂ signals measured underneath the cells (red ROI, white arrows) compared with their immediate surroundings (green ROI), suggesting that the plated myocytes contribute to the formation of an unstirred layer with a significant diffusion delays. Such delays were seen not only in association with the rounded, irreversibly contracted cells that dominated the scene after several hours of experimentation but also under the intact cells (lower row of images in Fig. 2) that were used in voltage clamp experiments.

Suppression of the cardiac Ca²⁺ channel by acute hypoxia.

We have previously reported that the current through cardiac Ca²⁺ channels is suppressed by ∼10% within 5 s and by ∼20–30%, depending on the charge carrier, within 45–60 s of exposure to hypoxic solutions (23). To gain further mechanistic insight, we carried out simultaneous measurements of O₂ withdrawal-induced rise in fluorescence signal and I_Ba in patch-clamped myocytes. I_Ba was activated three times, at 5 s intervals by voltage clamp depolarizations from −50 to 0 mV (Fig. 3A), with the middle pulse occurring while the cell was experiencing hypoxia as evidenced by the O₂ signal. The switch from fully oxygenated to hypoxic solution was digitally triggered and allowed to continue 100 ms after the depolarization had ended, assuring that changes in flow pressure did not affect I_Ba (6, 17). Fig. 3B shows sample I_Ba traces where a, b, and c correspond to identically labeled peak values of I_Ba in Fig. 3A at times before, during, and after the exposure to hypoxia. It is apparent that the peak I_Ba was suppressed by ∼20% within 2.5 s of hypoxia and that the suppression of the current at the end of the 100 ms depolarization pulse was proportionally larger. The recovery of current 5 s after reintroduction of O₂ was nearly complete. By simultaneously measuring dissolved O₂ tension and I_Ba, we found that a rapid (τ_{1/2 =} 225 ms), consistent, and accurately timed change of solutions could be achieved (Fig. 3C). However, this result was only secured by placing the puffer close to the voltage-clamped cell and by consistently flushing the tubing which carried the N₂ and O₂ perfused solutions so that a homogeneous oxygen tension was maintained. Without such precautions or with imprecise placement of the solution puffer, the measured paracellular O₂ concentration was found to change more slowly and the steady-state signals were more variable (Fig. 3D). In the absence of consistent, rapidly changing O₂ signals, it would be impossible to assess very short durations of hypoxia (<1 s) as a steady state is never reached. The O₂ signals in Fig. 3, C and D, were generated by a sequence of hypoxic solution pulses that increased in duration in increments of 500 ms, shown schematically in Fig. 3E (cf. T in Fig. 3A).

Fig. 3.

Simultaneous measurements of O₂ signal and Ba²⁺ current through the Ca²⁺ channel. A: standard voltage clamp protocol with 5 s pulse intervals and 30 ms depolarizations from −50 to 0 mV. All traces in figure are low pass filtered at 500 Hz. B: overlay of I_Ba from single sweep of protocol with 2.5 s exposure to N₂-bubbled solution. Letters a, b, and c correspond to labeled depolarizations from A. C: example of rapid changes in O₂ tension measured in the paracellular vicinity of a voltage clamped cell. Multiple overlaid sweeps are shown with exposure to hypoxia ranging from 0.5 to 2.5 s. D: example of slower and less reproducible changes in Po₂ measured in the paracellular vicinity of a voltage-clamped cell under less optimal conditions. E: timing of the command potentials to the puffer (top) and voltage-clamp apparatus (V_m, bottom).

While monitoring O₂ tension to assure that solutions were being rapidly exchanged, we found that short durations of acute hypoxia transiently diminished l-type Ca²⁺ current carried by I_Ba. Fig. 4A shows sample experimental and control traces from myocytes that were exposed to puffs of 2.5 s duration prior to the second depolarization pulse with solutions equilibrated either with N₂ or O₂, respectively. When myocytes were subjected to the puffs of N₂ equilibrated solutions, the cell experienced a 6% decrease in I_Ba that recovered partially after 5 s, as compared with no change in current amplitude or kinetics when the cell was perfused with only oxygen-containing solutions. Fig. 4B shows the peak current density from an experiment (black and white filled circles) plotted against time, with durations of hypoxia ranging from 2.5 to 0.5 s. The hollow points, representing depolarizations during which time the cell experienced N₂, show that short durations of hypoxia suppress I_Ba, but the effect is smaller with shorter durations of N₂ equilibrated solutions. We found that durations of hypoxia equal to or longer than 1.5 s consistently and transiently suppressed I_Ba. There was always significant rundown of the current (black curve) even in presence of Ba²⁺ as the charge carrier. Rundown is the slow decrease of measured ionic current as a function of time, likely through the loss of molecules from within the cytoplasm that modulate channel function (11). These findings also suggest that the recovery of I_Ba after reintroduction of fully oxygenated solution was ∼80–90% complete after 5 s.

Fig. 5 shows the average suppression of I_Ba activated with 30 ms depolarizations as cardiomyocytes were exposed to hypoxia for 0.5, 1, 1.5, 2, and 2.5 s. The suppression of I_Ba increases in a linear fashion with increasing durations of hypoxia, with the strongest suppression of 10% occurring with 2.5 s of hypoxia exposures. In sharp contrast, when control myocytes were subjected to the switch of identical fully oxygenated solutions there were no significant changes in I_Ba.

Fig. 5.

Rapid suppression of I_Ba at the onset of acute hypoxia. Mean suppression of I_Ba by durations of hypoxia ranging from 0.5 to 2.5 s (black). Mean control suppression (diagonal lines) represent experiments in which the cells were alternatively puffed with two O₂ perfused solutions.

DISCUSSION

Here we have demonstrated a method to measure rapid changes in O₂ tension in the immediate vicinity of individual cells (τ_1/2 = 225 ms) and have used this approach to show that the hypoxic suppression of the current through the cardiac Ca²⁺ channel takes effect within 1.5 s of hypoxic insult and is rapidly reversible. The findings are relevant to identifying the mechanisms for rapid O₂ sensing in biological tissues and for evaluating the efficiency of rapid perfusion techniques and the speed with which the concentration of O₂ and other compounds can be altered at the surface of single isolated cells.

Hypoxic suppression of the cardiac Ca^{2 ±} channel.

We have previously reported that that I_Ca in cardiac cells is suppressed by ∼20% within 50 s of hypoxia and have argued that the speed of this response was too fast to reflect a significant buildup of reactive oxygen species or change in the ATP/ADP balance (23). Like other investigators of acute hypoxia (9, 21, 25, 28), we previously puffed deoxygenated solution from reservoirs bubbled with 100% N₂ onto voltage-clamped cells and collected electrophysiological data without directly measuring the local paracellular Po₂. The work reported here was undertaken to determine directly how rapidly the suppression of the cardiac Ca²⁺ channel follows the reduction in Po₂. As shown in Figs. 4 and 5, the suppression of the peak I_Ba through the Ca²⁺ channel developed linearly with time, reaching a 10% suppressive effect after 2.5 s exposure to N₂-bubbled solution. In some cells the suppression was considerably larger, especially with regard to the maintained component of I_Ba (Fig. 3B). We have observed that the suppression depends on the permeating ion, the channel phosphorylation state, and isoform of the channel (23), although there may yet be other factors that accentuate or modulate hypoxic regulation. Control experiments, where we switched back and forth between different fully oxygenated solutions (Figs. 4 and 5), suggest that the observed suppression was due to hypoxia and not to artifacts such as mechanical shock waves generated by the valves of the electromagnetically controlled solution puffing system (6, 24). Considering the gradual rundown of I_Ba (Fig. 4B), present also in the control experiments and those where no solution exchange took place, the full recovery of hypoxic suppression on reoxygenation suggests specificity to I_Ca suppressive effect (Figs. 3B and 4B).

The speed and reversibility of the measured hypoxic suppression of I_Ba suggest that the effect of O₂ on the Ca²⁺ channel is fairly direct and is not mediated by a slowly activating signaling process such as phosphorylation or alterations in concentrations of cellular constituents such as ROS, ADP, and ATP. This begs the question whether the heme-oxygenases that have been implicated in O₂ sensing in the carotid body (19, 27) and ganglionic nociceptive neurons (7) might also govern the hypoxic regulation of the cardiac Ca²⁺ channel. Published experiments with inhibitors of heme-oxygenases protoporphyrins are consistent with this possibility (23), but it remains to be determined to what extent heme-oxygenases may serve as a general extramitochondrial O₂ sensor. Alternatively, hypoxia may induce ROS signals that locally occur faster than with the delay of ∼3 min found with global cellular ROS measurements (13). This possibility is supported by reports of stretch-activated ROS production by sarcolemmal NADPH oxidase (22) and spontaneous mitochondrial superoxide flashes (26), both of which take place on a timescale of seconds or faster, but are not known to do so in response to hypoxia. The methods described here may be used in future experiments to distinguish between these possibilities by comparing the time course of the changes in Po₂ and I_Ca to local ROS signals or responses mediated by heme-oxygenase.

Verification of rapid changes in paracellular Po₂.

The use of O₂-sensitive coverslips allowed for the measurement of Po₂ directly under and around cardiomyocytes. Cells were adhered to a thin, inert coating consisting of a fluorescent ruthenium-based sensor imbedded in a matrix of organically modified silicates. In our implementation we used coverslips as a light guide for continuous evanescent wave excitation and measured fluorescence intensity. Epi-illumination, as used in the imaging studies, produced some autofluorescence from the polyethylene outflow tube of the solution puffer, but the responses were not significantly degraded. Similarly, instead of fluorescence intensity, it may do as well to measure fluorescence lifetimes (15). In either case the O₂-dependent signal decreases nearly linearly with Po₂, which in turn implies that that it may be rather difficult to measure very small values of Po₂ (<5 mmHg) without using a more sensitive sensor (with a larger Stern-Volmer constant). On the other hand, for our present purpose the kinetics of the measurements were of greater importance. By ratiometric imaging of the Po₂-dependent fluorescence we found that transitions between hypoxic and fully oxygenated solution occurred rapidly (τ_1/2 = 225 ms) only right in front of, and close to, the outflow tube of the solution changer. It appeared that the stream of solution, which hit the coverslip at an incident angle of ∼45°, rapidly fanned out to cover a larger area and thereby slowed down considerably at the sides and with increasing distance from the exit port. It is known that much faster transition times (<5 ms) can be achieved in electrophysiological measurements where a small voltage-clamped HEK-295 cell is lifted off the coverslip and placed in one of the streams from a piezo-electric solution changer (1), but such procedures are hardly practical with the larger ventricular cardiomyocytes (∼100 μm × 20 μm × 10 μm). Furthermore Fig. 2 shows that the Po₂ signal under the attached cells were delayed by ∼50 ms, suggesting a diffusion (D ≅ 2μm²/ms) delay through an unstirred layer comparable to the thickness of the cell. Therefore the value of ∼50 ms may represent a practical limit for the speed of application of O₂ and other compounds that must penetrate the intact cells before they take full effect. The measurements of Po₂ suggest that it is important to verify the speed and effectiveness of solution puffers and that the electromagnetically controlled puffer used in this study was adequate to implement changes in Po₂ well before the gradual hypoxic suppression of the Ca²⁺ channel reached a significant level after about 1.5 s. The technique described here may prove useful in testing dose-dependence and kinetics of various hypoxic responses in different cell types.

GRANTS

This work was supported by National Heart, Lung, and Blood Insititute Grant RO1-HL107600 (M. Morad).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: J.A.S. and L.C. performed experiments; J.A.S., A.O.R., M.M., and L.C. analyzed data; J.A.S., A.O.R., M.M., and L.C. interpreted results of experiments; J.A.S. and L.C. prepared figures; J.A.S., A.O.R., M.M., and L.C. drafted manuscript; J.A.S., A.O.R., M.M., and L.C. edited and revised manuscript; J.A.S., A.O.R., M.M., and L.C. approved final version of manuscript; M.M. and L.C. conception and design of research.