Abstract Guinea pig hearts were stained with a voltage-sensitive dye and imaged on a photodiode array to record fluorescent action potentials (APs) from 124 sites. Activation and repolarization patterns were recorded from the epicardium during stimulation at different loci and correlated with the underlying fiber architecture. Endocardial APs were recorded by inserting a light guide into the ventricular cavity or by dissecting out the ventricular free wall to expose the endocardium. In hearts paced on the right atrium to simulate sinus rhythm, activation emerged synchronously over a large area of the ventricular epicardium and spread laterally in 5 to 7 ms. The apparent longitudinal and transverse velocities were 2.66±0.11 and 1.65±0.09 m/s (n=12). In contrast, repolarization began near the apex on the endocardium and spread transmurally in 6±1.3 ms (n=12) and then anisotropically along the epicardium in 25 to 30 ms with apparent maximum (0.53±0.11 m/s) and minimum (0.31±0.10 m/s) repolarization velocities that aligned with the longitudinal and transverse axes of epicardial fibers. When paced on the epicardium, activation of intact hearts (n=12) and perfused sheets (n=8) was anisotropic, with longitudinal (0.85±0.05 m/s) and transverse (0.44±0.04 m/s) conduction velocities that aligned with the epicardial fiber orientation. When activation was initiated at different sites on the epicardium, repolarization always began near the apex and exhibited patterns similar to those obtained under right atrial pacing, but with slower longitudinal (0.41±0.09 m/s) and transverse (0.23±0.07 m/s) repolarization velocities (n=18). In sheets stretched parallel to the longitudinal axis of surface fibers, AP durations (APDs) increased as a function of fiber length, from the length at zero developed tension to 120% of the length at maximum developed tension (Lmax). Spatial distributions of APDs did not change during stretches along the rising phase of the length-tension curve. In sheets stretched to 50% of Lmax, APDs were shorter and more homogeneous on the endocardium (mean APD, 188 ms; ΔAPD, 195−186=9 ms) than on the epicardium (mean APD, 204 ms; ΔAPD, 212−186=26 ms; n=8). In guinea pig hearts, activation is rapid; therefore, repolarization depends primarily on intrinsic spatial heterogeneities of APDs. Consequently, repolarization begins at endocardial cells with the shortest APDs and spreads transmurally and then anisotropically on the surface according to the epicardial cell orientation.
- myocardial fiber orientation
- action potential duration
- optical mapping
- repolarization patterns
- activation patterns
Activation and repolarization patterns in mammalian hearts are important in understanding the mechanisms of normal and abnormal cardiac rhythms. However, although activation has been extensively characterized by use of extracellular electrodes,1 repolarization remains incompletely understood because of the poor resolution of T waves measured by surface electrograms. Measurements of repolarization to assess spatial heterogeneities of APDs have relied on (1) potential gradients,2 3 (2) refractory period determinations,4 5 (3) monophasic APs,6 (4) ARIs,7 8 or (5) intracellular microelectrodes9 by using isolated sheets of ventricular tissue, papillary muscles, or enzymatically dispersed myocytes. Each method offers advantages but also has limitations.
Potential gradients measured with 120 unipolar electrodes on the ventricles have suggested that repolarization is spread anisotropically during epicardial stimulation. However, during endocardial or midmural stimulation, potential gradients did not convey a three-dimensional view of repolarization patterns.10 Refractory periods were determined by recording the time of reexcitation after the application of extrastimuli at the recording site, but the method is time consuming, since a set of premature stimuli must be tested for each site.4 ARIs use an array of extracellular electrodes. The activation time point at each site is detected at the maximum negative point of the QRS complex [(dV/dt)min] and the recovery time point at the maximum positive slope of the T wave [(dV/dt)max].8 ARIs correlate with APDs measured with intracellular electrodes and with refractory periods measured by programmed stimulation.7 8 ARIs predicted similar patterns for activation and repolarization in dog hearts.8 Regrettably, substantial errors can occur in the estimation of activation and recovery, particularly under nonideal conditions, including alterations in recording sites relative to the activation sequence, nonuniform coupling resistances, and anisotropic membrane properties.11 Contact (suction or pressure) electrodes record monophasic APs, which closely resemble APs recorded with intracellular electrodes, and have been used to study epicardial and endocardial repolarization in vitro and in vivo. But, at most, 8 to 10 contact electrodes can be simultaneously used, and there arc concerns regarding cell depolarization and tissue damage under the recording electrode.12 Monophasic APs recorded with a single contact electrode suggested that APDs were inversely related to activation time.6 Intracellular microelectrode recordings from tissue patches and enzymatically dissociated guinea pig myocytes indicated that APDs are shorter at the apex than the base of the left ventricle.9 However, a concern with this approach is that APDs are modified by the lack of stretch and might not have the same characteristics as cells in intact hearts. Furthermore, intracellular electrodes offer limited sampling because of the difficulty of maintaining them in place on a beating heart.
Voltage-sensitive dyes and imaging techniques have been previously used to measure APDs and repolarization patterns in “working” guinea pig hearts.13 14 However, these earlier studies were complicated by movement artifact.15 Since the upstrokes of APs always precede force generation, they are not subject to motion artifact that might distort activation patterns. On the other hand, the downstrokes of APs are in phase with peak myocardial force development and the onset of relaxation. The downstrokes of optical APs in frog ventricular muscle were readily measured in zero Ca2+ Ringer’s solution, which stopped contractile movement,14 but removal of extracellular Ca2+ prolongs APDs and may alter repolarization patterns and cause rhythm disturbances in mammalian hearts. Movement could also be abated by perfusing hearts with negative inotropic agents, such as 2,3-butanedione monoxime16 17 or verapamil,18 but the concentrations needed could alter upstroke velocities and APDs and interfere with mechanoelectrical feedback.19 20 21 Novel chamber designs and perfusion techniques now make it possible to place a Langendorff-perfused or a working heart in a chamber with a glass window or to pin a perfused working ventricular sheet on a rack to adjust muscle length and thus record APs essentially free of movement artifacts, without resorting to negative inotropic agents.22 23 24 25 26 Newer dyes like di-4-ANEPPS27 28 and RH-42129 also offer improved signal-to-noise ratios by yielding larger fractional changes in fluorescence per unit change of membrane potential and by fluorescing at long wavelengths (650 to 750 nm), where the quantum efficiency of photodiodes is >90%. For optical APs that are distorted by mechanical artifacts (≤25% of the APs), repolarization time points can still be uniquely identified through the maximum second derivative [(d2F/dt2)max] of the fluorescence downstroke.22 23 24 In Beeler-Reuter simulations of the ventricular AP and in experiments on guinea pig myocytes and intact perfused hearts, (d2F/dt2)max was found to fall at 97±2% of recovery to baseline.30 The coincidence of (d2F/dt2)max and AP repolarization was maintained during changes in APD elicited by frequency, hypoxia, and ischemia.30 Moreover, under physiological conditions at normal resting potential, (d2F/dt2)max was shown to be coincident to the refractory period and could be used to map repolarization and refractoriness in a single heart beat.26 The close correlation between optically recorded APs and those recorded with an intracellular microelectrode has been well documented.15
In the present study, activation and repolarization waves from the epicardium and endocardium of guinea pig hearts were measured under different stimulation protocols by using a voltage-sensitive dye (RH-421), a light guide, and a photodiode array. The guinea pig heart offers several advantages: (1) The size of the heart made it possible to image electrical activity and analyze fiber architecture (through serial sectioning) from the entire left or right ventricular free wall. (2) The excellent collateral circulation made it possible to develop an isolated ventricular free wall preparation perfused and superfused in blood-free oxygenated Ringer’s solution for several hours with no significant metabolic deficiencies. (3) The AP has been well characterized, with a long APD and a classical APD versus frequency relation. Unlike other small rodents, the electrogram waveform of the guinea pig heart is similar to that in the human heart, even though the guinea pig AP is devoid of Ito.30
Materials and Methods
Intact Heart Preparation
Guinea pigs (350 to 450 g, either sex) were anesthetized with sodium pentobarbital (30 mg/kg IP), and their hearts were rapidly excised, cannulated at the aorta, and retrogradely perfused in a modified Langendorff setup. A key feature of Langendorff preparations is that the hearts do not perform mechanical work because of the absence of left ventricular filling and diastolic fiber stretch. This ensured that spatial heterogeneities of fiber stretch did not alter APDs to produce nonuniformities of APDs. The hearts were then stained with a voltage-sensitive dye (RH-421) and placed in a fluid-filled Plexiglas chamber with an optical window for monitoring fluorescence as previously described.13 23 24
Several bipolar stimulating electrodes (250-μm-diameter polytetrafluoroethylene [Teflon]–coated Ag+/AgCl wires with a 50 μm interelectrode distance) were simultaneously positioned on the intact heart: (1) For right atrial pacing, an electrode was glued to the right atrium with cyanoacrylate adhesive. (2) For endocardial pacing, an electrode was inserted into the ventricular chamber and implanted on the endocardium. (3) For epicardial pacing, one to five stimulating electrodes could be simultaneously positioned at the apex, base, anterior, posterior, or center of the epicardium.
Ventricular Sheet Preparation
The left ventricular free wall was isolated with its coronary perfusion intact by excising the right ventricle, the septum, and left papillary muscles from a Langendorff-perfused heart. The removal of a ventricle and the septum decreased perfusion pressure. The ligation of three to five cut coronary artery branches supplying the excised tissue raised the perfusion pressure above its initial value of 80 mm Hg, which was reestablished by decreasing the flow rate. Typically, if two thirds of the tissue mass was removed, the flow rate needed to be reduced by approximately two thirds.
Perfused sheets were pinned horizontally on a two-piece Delrin platform (Fig 1A⇓). One half of the platform was fixed, whereas the other could slide on a set of stainless steel tracks. The movable platform was connected to a tension transducer (Gould, model M602) mounted on a micromanipulator. Heated Ringer’s solution (35±1°C) was delivered to both the aortic cannula and the tissue chamber inlet to respectively perfuse and superfuse the preparation. Ventricular free walls were pinned with either the epicardium or endocardium facing up, with up to five sets of bipolar stimulating electrodes (250-μm-diameter Teflon-coated Ag+/AgCl with a 50 μm interelectrode distance) placed on the top surface and one set placed on the bottom surface. A No. 2 glass coverslip (22×22 mm) was placed at the fluid-to-air interface above the tissue to prevent fluid vibrational noise during optical recordings. Ventricular sheets were passively stretched by displacing the movable section of the rack by increments of 200, 400, or 800 μm. Equilibrium length (Lo) was defined as the distance between the fixed and movable pins when further stretches registered an increase in passive and developed tension as measured by the transducer. Stretches produced systematic increases in passive and developed tension until developed tension reached a maximum, at length Lmax. The length-tension curves indicated that Lmax was 30% to 38% greater than Lo. The coronary perfusion of isolated ventricular sheets was of critical importance when attempting to measure APs with negligible motion artifacts and to avoid tissue ischemia, which decreased APDs and enhanced spatial inhomogeneities of APDs.
Intact hearts were illuminated with light from a 200-W mercury-xenon arc lamp (PTI, Fig 1B⇑), which was collimated with a parabolic reflector and cooled with a heat filter (water-filled 4-in optical path length). The excitation beam was controlled with an electronic shutter (Melles Griot) to illuminate the heart only during optical recordings. The beam was passed through an interference filter (520±20 nm, Omega Optical), reflected by a 45° dichroic mirror (600-nm long-wave pass, Corion), and focused on the stained heart with an epi-illumination lens (50 mm, 1:1.8 E series, Nikon). Epifluorescence was collected and passed through a 645-nm cutoff filter (Schott Glass, RG-645) and focused onto a 12×12 photodiode array (Centronics). The same optical apparatus was also used to record signals from myocardial sheets, but the optical axis (normal to the surface of the tissue and the photodiode array) was vertically oriented. In this instance, light from two 100-W tungsten-halogen lamps (PTI) was collimated, passed through 520±20-nm interference filters (Omega Optical), and refocused on the preparation at a 45° angle of incidence. Epifluorescence from stained tissue was collected, passed through a 645-nm cutoff filter, and focused on the array. The physical dimensions of the photodiode array were 18×18 mm (1.4×1.4 mm per diode with 0.1 mm of dead space between diodes), but the area of the tissue image that was focused on each array diode was either 1×1 mm (intact hearts) or 0.5×0.5 mm (isolated sheets), with a depth of field of 174 and 44 μm, respectively.
Simultaneous APs could also be recorded from a 1-mm-diameter region of the endocardium by using a bifurcated light guide (Dolan-Jenner Industries) and from the epicardium by using a photodiode array (Fig 1B⇑). One branch of the guide focused an excitation beam on the endocardium; the second branch collected and transmitted focused fluorescence emission to a photodiode (EG&G Reticon). The source of the illumination was light from a 45-W tungsten-halogen lamp (PTI) that was collimated and passed through an interference filter (520±20 nm, Omega Optical). The tip of the light guide, which was bent at 90° at 1 mm from its extremity, was inserted into the left ventricular chamber through the mitral valve. Light from the fiberoptics was detectable across the ventricular wall, permitting the guide to be aligned opposite a desired epicardial region of tissue monitored by one of the photodiodes on the array.
Intact hearts and isolated ventricular free walls were typically paced at a cycle length of 300 ms (≈200 bpm) with square pulses of 1-ms duration at twice diastolic threshold (twice capture voltage). This rate was within the physiological range for guinea pigs (200 to 300 bpm at 38°C). A constant SA nodal rhythm was obtained by pacing the right atrium of intact hearts near the SA node at 200 bpm. Experiments were carried out at a slightly lower temperature (35±1°C) to be able to pace the heart reliably and overcome the intrinsic rate of the pacemaker. There were no observable differences in activation or repolarization patterns when hearts were maintained at 38°C. However, APDs uniformly shortened with increasing temperature. Stimulation of the endocardium was used as an alternative method to activate the ventricles through the specialized conduction system. Epicardial stimulation was used to avoid activation of the specialized conduction system and permitted measurements of AP spread in ventricular cells without the influence of Purkinje fibers. During endocardial and epicardial stimulation of intact hearts, the right atrium was simultaneously paced to prevent out-of-phase SA nodal impulses from capturing the ventricles and interfering with activation and repolarization patterns.
After electrophysiological recordings were made, the region of tissue viewed by the array was marked with fiducial points by impaling its edges with a microelectrode dipped in india ink. The markings were placed so as to correspond to the center of tissue patches viewed by diodes on the corners of the array. The ventricular free walls were excised and fixed in Bouin’s solution with their epicardial surfaces placed against a glass slide for 1 hour and then overnight without a slide. All tissues were in diastole before the fixation protocol. Fixation flat against a glass slide ensured that sections were cut parallel to the imaged plane of the heart for accurate correlation with maps of electrical activity. Fixed tissue was embedded in paraffin, serial sections were cut every 5 μm, and every 10th section was mounted on a glass slide. To ensure an accurate measurement of the depth of all sections relative to the surface, care was taken to identify the front surface of the epicardium. Mounted sections were progressively stained with Mayer’s hematoxylin, counterstained with eosin, and placed under a glass coverslip.31
Stained sections were examined under a microscope between cross polarizers to better visualize fiber orientation. An image of each section was projected, magnified, and refocused on a 12×12 reticule, which conformed to the dimensions of the array and thus to the tissue before fixation. Shrinkage caused by fixation (typically 35%) was automatically taken into account by aligning the fiducial marks on the sections with their respective locations on the reticule. The fiducial marks were also used to align the sections with respect to orthogonal cartesian coordinate axes, where the base to apex axis was parallel to the vertical y axis. The orientation of fibers viewed by each photodiode was measured with a protractor mounted in an eyepiece. One hundred twenty-four fiber orientations were averaged to give the longitudinal fiber axis for the tissue imaged by the array. The average fiber orientation (αF) of each section was plotted as the function of its depth in the right and left ventricular walls.
Tests and Precautions
APDs depend on physiological parameters that include heart rate, temperature, resting membrane potential, and the metabolic and contractile states of the myocardium. Consequently, experimental conditions were controlled to ensure valid measurements of spatial heterogeneities of APDs. Heart rate was controlled by pacing the preparations and continuously monitoring the cycle length to detect possible rhythm disturbances. We verified that spatial heterogeneities of temperature were negligible by measuring the temperature at various sites on the surface of the epicardium with a 1-mm-diameter thermistor to verify that it was within ±1°C of the set point (35°C). The thermistor was inserted between the heart and the glass window of the chamber and was systematically displaced along the heart’s surface. Heat from the excitation beam raised the temperature <1°C in 60 s and did not cause spatial variations of temperature. Temperature effects caused by the excitation beam were not significant, since the preparations were illuminated only during data acquisition (10 to 15 s). Fluid accumulation and pressure in the ventricular cavity could stretch the left ventricle and thus alter APDs. This concern was addressed by cutting the mitral valve to relieve fluid buildup in the ventricular cavity. Mechanical contact and pressure between the window of the chamber and the epicardial surface did not alter APDs or cardiac function. This issue was verified by measuring optical APs in the absence and presence of the chamber and with increasing pressure between the front and the rear portion of the chamber to compress the heart. The compression of a 20-mm-diameter heart by 1, 2, 3, and 4 mm did not alter coronary flow rate, pulse pressure, surface electrogram recordings, or APDs. However, it increased diastolic pressure respectively by 0, 2, 5, and 15 mm Hg. Further compression of 5 and 6 mm increased diastolic pressures by 30 and 50 mm Hg, reduced coronary flow, altered electrogram recordings, and shortened APDs. The present experiments were carried out with 2- to 3-mm compression, which reduced motion artifact without changing AP characteristics, coronary flow rates, or pulse pressure. Similar spatial heterogeneities of APDs were also measured in isolated ventricular sheets stretched along the rising phase of their length-tension curve. Thus, the use of the chamber to restrain movement did not modify activation and repolarization patterns.
To visualize patterns of activation and repolarization, maps were displayed as isochrones on the basis of gray scales. These maps were generated from the activation or repolarization times of n groups of the 124 APs detected by the photodiode array. Each group was assigned a gray tone, where light to dark shades represented early to late activation or repolarization times. Each diode location (box) on a symbolic map of the array was then filled with the appropriate gray tone representing that channel’s activation or repolarization time group. Five gray tones allowed for the best visualization of the overall pattern while maintaining an excellent degree of detail.
Fig 4⇓ shows the anatomic landmarks of the guinea pig heart superimposed on a schematic of the array to depict the regions of the left (A) and right (B) ventricles that were examined. The same regions of the left and right ventricles were examined throughout this study. In Fig 2B⇓, activation time delays were used to construct gray-scale isochronal maps during the delivery of stimuli to the center of the left ventricular epicardium. With stimuli delivered every 350 ms, a 1.2-s scan recorded three cardiac beats. Three representative APs are shown in Fig 2⇓ from early (1), intermediate (2), and late (3) regions of tissue activation; the fourth trace is a bipolar electrode recording. The complexity of the electrogram waveform emphasizes the difficulty of defining activation and repolarization time points from surface electrode recordings. Gray-scale activation maps from three successive beats (A through C) exhibited similar anisotropic patterns of propagation, with activation times across the tissue of 19.2 (A) and 18.4 (B and C) ms, which were reproducible over 2 to 3 hours (data not shown). The spread of APs was anisotropic and produced elliptical isochrones, where the major and minor axes of the ellipse represented the directions of fastest and slowest conduction velocities.
Detection of Repolarization Time Points
The time of depolarization has been routinely defined and detected through the maximum first derivative of the AP upstroke, (dV/dt)max, measured with intracellular microelectrodes. For repolarization, there is no simple definition nor standard algorithm. For example, values of 50%, 75%, and 90% recovery of the downstroke to baseline have each been used as definitions of the repolarization time point. With optical recordings, the same approaches can be used to define activation and repolarization by signal processing of the voltage-dependent fluorescence signals. However, for optical recording, the recovery to baseline is more susceptible to movement artifacts that alter the kinetics of AP downstrokes. Recently, the peak of the second derivative was shown to reliably detect the repolarization time point of optical APs,22 23 24 because the downstroke of APs possesses an inflection point that can be detected despite the presence of motion artifacts. In Fig 3⇓, APs and phasic contractions were simultaneously recorded from a 6×6-mm region of a perfused left ventricular sheet stretched to 50% of Lmax. One AP was chosen for its lack of movement artifact (diode 52), and the next three APs were chosen because movement artifacts altered their downstrokes in different manners (diodes 80, 81, and 124). The second derivatives, (d2F/dt2)max values, of these tracings (52″, 80″, 81″, and 124″) are shown along with simultaneous recordings with a bipolar electrode, developed tension, and its second derivative (BE, T, and T″, respectively, in Fig 3⇓). Despite the different AP downstrokes caused by movement artifacts, the maximum second derivatives of AP downstrokes occurred at 97±2% of return to baseline, which identified a set of unique repolarization time points. APDs determined from the time interval between (dF/dt)max and (d2F/dt2)max predicted reproducible APDs (±1.4 ms) from beat to beat at each diode. Adjacent diodes (channels 80 and 81) had similar APDs, despite widely varying movement artifacts. Movement artifacts did not shift (d2F/dt2)max, most likely because their kinetics are similar to the phasic contractions (T), which had flat second derivatives during the repolarization phase of APs (T″). The use of the (d2F/dt2)max inflection point of AP downstrokes provided a practical and reliable method to map APDs and predicted the systematic spread of repolarization waves along the epicardium.
Fiber Orientation as a Function of Depth
Histological analysis was carried out after optical recordings in order to correlate ventricular fiber orientations at different depths in the tissue with activation and repolarization patterns recorded under various stimulation protocols. In hearts cannulated at the aorta and hanging freely, fiber orientation was measured with respect to arbitrary cartesian coordinates. The origin was at the center of the ventricular free wall, the ordinate was aligned vertically from base to apex, and the abscissa was aligned horizontally from anterior to posterior. The longitudinal fiber axis of left ventricles was oriented at a 135±5° (or 135°+180°=315°) angle on the epicardium and rotated counterclockwise by 135° from epicardium to endocardium where fibers aligned with the vertical axis (Fig 4⇓, inset A). The longitudinal axis of epicardial fibers on the right ventricles was at a 15±5° (or 15°+180°=195°) angle and rotated by 75° from epicardium to endocardium where fibers again aligned with the vertical axis (Fig 4⇓, inset B). Fig 4⇓ represented the average of histological studies from left (n=6) and right (n=6) ventricles. From serial sections taken at different depths, fibers on the right epicardium (0.8 mm thick) were at 15±5° for the first 150 μm, were rotated gradually counterclockwise by ≈10° per 100 μm for 500 μm and more abruptly by ≈22° for 50 μm, and were then aligned with the vertical axis for 100 μm of endocardial wall. In the thicker left ventricles (1.5 mm), epicardial fibers were oriented at 135±5° for the first 200 μm and rotated gradually counterclockwise at ≈5° per 130 μm for 650 μm and then more abruptly by ≈10° in 65 μm for 650 μm until the endocardial fibers aligned with the vertical axis. It is important to note that the data shown in Fig 4⇓ were obtained on fixed tissue, which should be corrected for ≈35% shrinkage. As a result, the actual thickness of the right ventricles was 1.08±0.14 mm (n=6) compared with 0.8 mm for the fixed tissue; left ventricles were 2.03±0.25 mm (n=6) compared with 1.5 mm.
Most noteworthy for both right and left ventricles was the vertical orientation of endocardial fibers and the consistent orientation of epicardial fibers at >250 μm from the surface. The orientation of epicardial fibers was uniform for a depth greater than the depth of field of the imaging system such that fiber rotation did not influence the present optical maps of electrical activation.
Superposition of Activation and Repolarization Patterns With Fiber Orientation
Perfused guinea pig hearts stained with RH-421 were placed in a fluid-filled chamber, and a 12×12-mm area of left or right epicardium was focused on the array, as shown in Fig 4⇑. In Fig 5⇓, activation (top panels) and repolarization (middle panels) patterns were recorded from the left ventricle during various stimulation protocols (A, B, and C) and correlated with fiber orientations at different depths of the tissue (bottom panels). During right atrial pacing near the SA node, activation broke through at the anterior epicardium and spread in 6.2 ms to the posterior epicardium (Fig 5A⇓). Activation spread anisotropically with the major axis parallel to the longitudinal axis of endocardial fibers (Fig 5D⇓). Repolarization measured from the same set of APs began near the apex and spread anisotropically across the tissue in 27.7 ms (Fig 5A⇓′). During epicardial pacing near the center (Fig 5B⇓) or upper left corner of the tissue viewed by the array (Fig 5C⇓), activation propagated anisotropically with a major axis oriented (at 135°) parallel to the longitudinal axis of epicardial fibers (Fig 5F⇓). Repolarization patterns elicited by right atrial pacing (Fig 5A⇓′) and epicardial stimulation near the center (Fig 5B⇓′) or upper corner (Fig 5C⇓′) of the array were also anisotropic, with major axes oriented 135° parallel to the average orientation (αF=135°, Fig 5F⇓) of the longitudinal axes of epicardial fibers. The major axis of repolarization patterns was within experimental error (±5°) parallel to the longitudinal axis of epicardial fiber (n=6).
In Fig 6⇓, a similar experiment was carried out on a perfused right ventricle. During right atrial and endocardial stimulation, the major isochrones for activation (Fig 6A⇓ and 6B⇓) were parallel to the long axis of endocardial fibers (Fig 6D⇓). During epicardial stimulation, the major axis of activation (Fig 6C⇓) was parallel to the long axis of epicardial fibers (Fig 6F⇓). Irrespective of the pacing protocol, repolarization spread anisotropically (Fig 6A⇓′, 6B′, and 6C′), with major axes at 30°, indicating a slight rotation of repolarization isochrones compared with the average longitudinal fiber axis at 15° to 20° (Fig 6F⇓).
Analyses of repolarization patterns measured from left and right epicardia indicated that repolarization was not random but traveled systematically across the heart’s surface. When measured during epicardial pacing, repolarization traveled anisotropically, with maximum and minimum velocities of 0.41±0.09 and 0.23±0.07 m/s (n=12) that aligned with the longitudinal and transverse axes of epicardial fibers, respectively. The repolarization patterns were similar to activation maps measured under the same pacing protocol but had slower repolarization velocities compared with 0.85±0.05 and 0.44±0.04 m/s for activation. These velocities represented intrinsic myocardial fiber velocities, since activation and repolarization patterns were guided by the orientation of surface fibers and had similar eccentricities (ratios of longitudinal to transverse velocities) of 1.93 and 1.78, respectively. During right atrial or endocardial pacing, activation of the epicardium was driven by the specialized conductile system (Purkinje fibers), resulting in maximum and minimum apparent velocities of 2.66±0.11 and 1.65±0.09 m/s, which aligned with the longitudinal and transverse axes of endocardial fibers, respectively. In contrast, repolarization first broke through on the epicardial surface from the apex (bottom of the array) to one third of the way up the wall of the ventricle (Figs 5A′ and 6A′) and spread with maximum and minimum apparent velocities (0.53±0.11 and 0.31±0.1 m/s, respectively; n=12), which remained predominantly aligned with the longitudinal and transverse axes of epicardial fibers.
The location of the first sites to repolarize did not significantly vary with changes in the stimulation site. During epicardial stimulation at the center of the left ventricle (Fig 5B⇑), the first sites to repolarize were typically located one third of the distance up from the apex (Fig 5B⇑′). When the stimulus site was shifted to the anterior corner, near the base (Fig 5C⇑), the first site to repolarize shifted slightly up, toward the base (Fig 5C⇑′). With all three stimulation protocols (A through C), APDs were shorter at the apex than the base (n=12). Note that in Fig 5⇑, APs near the apex (Fig 5C⇑, bottom channels) depolarized after APs near the base (Fig 5C⇑, top right), yet the apex (Fig 5C⇑′, bottom) repolarized before the base (Fig 5C⇑′, top right). Under right atrial (Figs 5A⇑ and 6A⇑) or endocardial (Fig 6B⇑) stimulation, activation emerged at the center of the epicardium, guided by the endocardial fiber orientation that had faster conduction velocities compared with those measured during direct epicardial stimulation. The orientations of maximum and minimum velocities during atrial pacing were aligned with the axes of endocardial fibers, which are activated by Purkinje fibers lining the endocardial surface. Thus, Purkinje-driven activation begins through the rapid activation of the endocardium, followed by transmural propagation that emerges to synchronously activate a large patch of epicardium.
To determine whether repolarization began on the epicardium or traveled to the surface from deeper cells, APs were simultaneously recorded from opposite sides of the ventricular wall. A light guide was inserted into the left ventricular chamber to record an endocardial AP while recording 124 APs from the epicardium by using the array (see “Materials and Methods”). During right atrial pacing, endocardial APs depolarized and repolarized before the epicardial APs immediately opposite them on the other side of the wall (Fig 7A⇓ and 7A⇓′), which indicated that repolarization traveled from endocardium to epicardium in 6±1.3 ms (n=6). When paced on the epicardium, the epicardium depolarized before the endocardium, but the endocardium still repolarized either at the same time (Fig 7B⇓ and 7B⇓′, n=4) or before the epicardium (not shown, n=4). During right atrial pacing, activation on the endocardium preceded that of the epicardium by 3 to 5 ms (n=8).
For the 1.6- to 1.9-mm-thick ventricles, the average transmural activation velocity was 0.48±0.12 m/s (n=8), which is similar to transverse activation velocities seen on the epicardium (0.44±0.04 m/s). Transmural activation velocities were similar (within experimental error) when APs spread from endocardium to epicardium (the right atrium was paced) or epicardium to endocardium (the epicardium was paced). Since it was assumed that repolarization began on the endocardium and traveled across the thickness of the ventricles (1.6 to 1.9 mm), the average transmural repolarization velocity was calculated to be 0.27±0.04 m/s (n=8), a value close to the transverse repolarization velocity of 0.23±0.07 m/s.
The light guide made it possible to measure endocardial APs from intact hearts and (if a direct path from endocardium to epicardium is assumed) provided an estimate of transmural activation and repolarization velocities. Even though the transmural activation and repolarization time delays were brief, these values were reproducible and could be resolved, given the sampling rate of the data (0.7 ms per frame). Moreover, transmural delays were in line with our measurements of transverse activation and repolarization velocities, given the width of the guinea pig ventricular wall. However, the light-guide approach was limited to one endocardial AP and could not resolve the two-dimensional spread of activity on the inner surface. Consequently, ventricular walls were dissected and stretched as sheets to expose the inner and outer surfaces of the same heart to optical array recordings.
Electromechanical Properties of Ventricular Sheets
In isolated tissue experiments, the image magnification was adjusted such that the array viewed 6×6- instead of 12×12-mm of tissue. The recording area was reduced because of the smaller area of endocardium and in order to avoid muscle injury at the sheet edges. Perfused sheets were stained with RH-421 and mounted on the horizontal chamber (Fig 1A⇑) to record APs at a known contractile state along the length-tension curve of the myocardium. Left ventricular sheets were mounted horizontally to view the epicardium and were rotated to stretch the muscle parallel to the longitudinal axis of the surface fibers (Fig 8⇓). The resting length (Lo) was determined by stretching the tissue until the transducer detected phasic contractions elicited by repetitive stimulation (cycle length, 350 ms). The tissue was stretched by 800-μm increments from Lo and allowed to equilibrate for 5 to 10 minutes to ensure stable contractions before recording APs (Fig 8b⇓). Stretches produced the expected increase in phasic contractions until Lmax (length at which maximum active tension is generated) was reached; further stretches reduced active tension. Length-tension curves were measured from three different preparations with 0.5 (Fig 8a⇓), 1.5 (Fig 8b⇓), or 2.5 (Fig 8c⇓) mmol/L Ca2+ in the perfusate, and AP maps were recorded as a function of stretch.
In Fig 9A⇓ and 9B⇓, the quality of the signals from the epicardium are shown for sheets exhibiting weak (0.5 mmol/L Ca2+ and 13.3% stretch) and strong (2.5 mmol/L Ca2+ at Lmax) contractions, respectively. Even when the tissue generated large phasic contractions (18 g tension) in 2.5 mmol/L Ca2+ at Lmax, AP upstrokes were readily detected, and repolarization phases were not obscured by mechanical artifacts. Some diodes detected distorted AP downstrokes (Fig 9B⇓; channels 13, 38, 61, 62, 73, 74, etc) because of motion artifacts. However, the signal processing technique described earlier identified distinct (d2F/dt2)max inflections and provided unambiguous repolarization time points for all except two APs (channels 73 and 74).
Length-tension relations and changes in APDs as a function of muscle length were analyzed for left ventricular sheets continuously paced at 350-ms cycle lengths. Fig 10A⇓ shows the average length-tension curve of 10 muscle sheets perfused with 1 mmol/L Ca2+ Ringer’s solution. The curves were normalized with respect to Lmax, which was typically 38% greater than Lo. When stretched 25% beyond Lmax, developed tension decreased by ≈70%.
Stretches to Lmax (or beyond) produced increases in APDs that did not reverse upon return of the muscle to its resting-length. This effect is illustrated in Fig 10B⇑, where APDs increased with increasing muscle length from Lo to 25% beyond Lmax and were held at 125% of Lmax for 30 minutes (open circle). When the sheets were slackened back to Lo, APs did not return to their original duration (170 ms) but remained at 194 ms. Stretching the muscle for a second time (solid circle) did not change APDs.
In Fig 10C⇑, APDs from four regions of the left ventricle were measured as a function of muscle stretch. The 124 APs recorded by the array were divided into four groups of APs from the base, apex, anterior (right side), and posterior (left side) regions. APDs increased as a function of stretch and leveled off at 25% beyond Lmax. APDs were longest at the base and anterior (right side) region of the ventricles and shortest on the posterior (left side) and apex regions. APDs from all four regions of the left ventricle increased in a stepwise manner as a function of muscle length, so that stretch did not significantly alter the regional differences in APDs. Increases in APDs as a function of stretch were measured with 0.5, 1.0, 1.5, or 2.5 mmol/L Ca2+ in the perfusate, and as expected, APDs increased with decreasing [Ca2+]o (Fig 10D⇑).
A major concern was that fibers at various depths of sheet were stretched to different extents and that at Lmax, fibers in the midwall or the endocardium may have become stretched beyond Lmax. To compare APDs from the endocardium and epicardium, care was taken to measure APs from both surfaces at similar levels of passive stretch and to avoid overstretching of the preparations. From a resting length (Lo) and under continuous pacing, the tissue was stretched by equal increments of length (100 to 200 μm), and Lmax was identified when a small stretch failed to elicit an increase in developed tension. The muscle was then slackened to 50% of Lmax and allowed to equilibrate for 10 to 20 minutes before recording APs.
Epicardial Versus Endocardial Repolarization Patterns
A sheet of perfused left ventricle was dissected, as described in “Materials and Methods.” Papillary muscles and free trabeculae were removed from the endocardium to reduce the curvature of the ventricle and thus obtain uniformly stretched fibers. The sheet was stained with RH-421 in the perfusate and then mounted on the horizontal chamber rack with the epicardium facing up and the longitudinal axis of surface fibers parallel to the direction of stretch. The sheet was stretched to 50% of Lmax and allowed to equilibrate during continuous pacing at a 350-ms cycle length before recording APs from the epicardium. In Fig 11⇓, activation (A through D) and repolarization (A′ through D′) patterns were measured from the epicardium during pacing at different locations (square pulse symbols) on the edges of the 6×6-mm ventricular surface. Activation began at the stimulus site and propagated anisotropically with the major axis parallel to the longitudinal axis of epicardial fibers (top panels). As the stimulus site varied, the first sites to depolarize changed accordingly. As in the intact heart, repolarization (bottom panels) began near the apex, regardless of the initial site of depolarization. During stimulation at the base or posterior edge of the epicardium, repolarization began at breakthrough sites near the apex of the anterior side of the left ventricle (lower right edge of Fig 11A⇓′ and 11B′). With stimulation at the apex or anterior edge of the sheet, repolarization also began near the apex but was initiated from both the anterior and posterior edges of the left ventricle (Fig 11C⇓′ and 11D′). Maps of APDs were similar to the repolarization patterns shown in Fig 11A⇓′ through 11D′ (not shown).
The same ventricular sheet was removed from the rack, flipped over so that signals could be recorded from the endocardium (Fig 12⇓), and restretched to 50% of Lmax, with the longitudinal axis of endocardial fibers parallel to the direction of stretch (90°/270°). Activation (top panels) and repolarization (bottom panels) maps were measured from 6×6 mm of endocardium while pacing at different edges of the endocardium. Stimulation at the anterior (Fig 12A⇓) or posterior (Fig 12D⇓) edges of the endocardium produced activation patterns with major axes aligned at 105°, approximately parallel to the endocardial fiber axis. However, pacing at the base (Fig 12A⇓) or apex (Fig 12C⇓) caused rotations of the activation patterns, with the major isochrones oriented at ≈180°. From heart to heart, the orientations of endocardial activation patterns were less predictable than those measured from the epicardium because of the tortuosity of free trabeculae lining the endocardium and because the layer of endocardial fibers rotated rapidly as a function of depth in the wall (10°/200 μm, Fig 4A⇑). In addition, pacing electrodes most likely stimulated several layers of fibers with different orientations. Nevertheless, endocardial repolarization was typically initiated near the first site to depolarize but did not travel in a systematic pattern and appeared to be random (Fig 12A⇓′ through 12D′). The distribution of APDs on the endocardium was random (not shown), unlike that measured on the epicardium.
Activation and repolarization patterns were measured from the endocardium and epicardium of 10 left ventricles. Conduction velocities were similar (within experimental error) on both sides of the tissue. Measurements were begun from the endocardial and then the epicardial surfaces, or vice versa, to avoid systematic errors in APDs caused by stretches during length-tension analysis. Measurements from both surfaces were completed in 60 to 90 minutes, which avoided loss of signals from dye washout, photobleaching, and/or rundown of the preparations. In 8 of 10 sheets, APDs were shorter on the endocardium than the epicardium of the same ventricular sheet. APDs varied on the endocardium from 186 to 195 ms, with a mean of 188 ms, and on the epicardium from 186 to 212 ms, with a mean of 204 ms. In 2 of 10 sheets, mechanical artifacts were too large to measure APDs from the endocardium.
Voltage-sensitive dyes and imaging and signal-processing techniques were combined to record activation and repolarization wave fronts from the epicardium and endocardium of intact guinea pig hearts and arterially perfused ventricular sheets. Although the present data were limited to two-dimensional surface recordings of electrical activity, patterns measured under different pacing protocols were correlated with the underlying fiber histology to formulate a three-dimensional conceptual view of the activation and repolarization process.
Fiber Orientation in Guinea Pig Hearts
Fiber orientation in guinea pig hearts has not been extensively examined. Our histological analysis of guinea pig hearts showed remarkable similarities to studies of dog hearts, despite major size differences. Early studies of dog hearts first fixed and then flattened the tissue and reported that left ventricular epicardial fibers were oriented at 120° instead of 135° and rotated counterclockwise by 115° instead of 135° from epicardium to endocardium.32 33 More recent studies reversed the procedure by first flattening and then fixing the tissue.34 35 This latter procedure was similar to the present study and reported results virtually identical to those described here. One study reported that right ventricular epicardial fibers were oriented at 20° instead of 15° and rotated counterclockwise by 70° instead of 75° from epicardium to endocardium.34 The other study reported right and left ventricular fiber orientations in dog hearts that were in excellent agreement with our guinea pig data.35 Besides differences in the manner in which earlier studies took into account the curvature of the tissue, they also examined a smaller percentage of the total ventricular area.32 33 In the present study, tissue architecture and electrical mapping were analyzed under similar conditions. Thus, it was necessary to flatten and then fix the tissue during diastole and to examine close to 100% of the free walls to reproduce the conditions during optical measurements. Others have shown that fiber orientations were similar during diastole and systole and thus independent of the contractile state of the tissue.32 33
Ventricular Repolarization Patterns
Physiological activation of the heart through pacing of the right atrium or pacing at various sites on the ventricles produced repolarization wave fronts that spread anisotropically along the epicardium of right and left ventricles according to the orientation of surface fibers (Figs 5⇑ and 6⇑). During epicardial pacing, the major and minor axes of the repolarization wave fronts traveled at velocities (0.41±0.09 and 0.23±0.07 m/s, respectively) that were twice as slow as activation velocities (0.85±0.05 and 0.44±0.04 m/s, respectively; Figs 5⇑ and 6⇑) and represented the intrinsic repolarization velocities for the myocardium. Repolarization patterns did not significantly change when the stimulation of the ventricles occurred through the specialized conductile system by pacing the atrium (Figs 5A⇑ and 6A⇑) or the endocardium (Fig 6B⇑) instead of direct stimulation of the epicardium. This demonstrates that the initiation of repolarization wave fronts was not significantly altered by the specialized conductile cells and originated at cells near the apex of the heart. During right atrial or endocardial pacing, there was a small but consistent increase in longitudinal and transverse repolarization velocities (0.53±0.11 and 0.31±0.1 m/s, respectively) compared with epicardial pacing. This apparent increase in repolarization velocity was attributed to the sequential appearance of multiple breakthrough sites of transmural repolarization that appeared on the epicardium. Attempts to shift the first site to repolarize away from the apex by pacing different regions of the myocardium were marginally effective in that the first site to repolarize was only slightly displaced by repositioning the stimulus electrodes (Figs 5B′, 5C′, and 11A′ through 11D′). These data indicated that groups of cells near the apex had shorter APDs compared with ventricular cells located elsewhere. Moreover, short APDs in cells near the apex were an intrinsic feature of these cells that was independent of the rate or the contractile state of the myocardium (altered by varying extracellular Ca2+).
Several lines of evidence suggested that repolarization of the epicardium was not a “phase” wave that appeared to spread because of the intrinsic repolarization of individual cells but was spread via intercellular electrical coupling, like activation. Precautions were taken to avoid spatial heterogeneities of myocardial fiber stretch and perfusion: (1) Repolarization spread anisotropically across the epicardium, generating elliptical isochronal lines with major and minor axes that were parallel to the longitudinal and transverse axes of the cardiac fibers. (2) Repolarization patterns exhibited highly reproducible orientations and velocities from beat to beat (not shown) and had different orientations on the right and left ventricles according to the underlying fiber orientations. (3) The slow spread of repolarization strongly suggested that repolarization velocities were driven by the rate of AP repolarization or −dV/dt of AP downstrokes by analogy with activation that is driven by dV/dt of AP upstrokes. (4) Stimulation of the heart by pacing the right atrium instead of the epicardium changed activation times and the orientation of activation patterns but did not change repolarization patterns. This implied that Purkinje fibers speed up ventricular activation but not repolarization. (5) Repolarization patterns were not due to systematic artifacts, since (a) random patterns could be observed as on the endocardium; (b) on the epicardium, repolarization could be reversibly transformed from an anisotropic spread to random repolarization events by hypoxic episodes25 ; and (c) although repolarization delays between adjacent diodes (2 to 3 ms) were close to the sampling rate (0.7 ms), these delays were significant because they progressed in a systematic and structured manner along the epicardium. Thus, the graded repolarization delays could not be interpreted as a phase wave that coincidentally appeared like spreading wave fronts but indicated that repolarization spread from cell to cell by a process analogous to activation.
Previous Models of Repolarization
There is no consensus to date regarding whether the repolarization of the myocardium actually spread from cell to cell or only appears to spread like a phase wave. Mathematical simulations were developed to model repolarization patterns, since they could not be resolved experimentally. The models were based on a repolarization process of the myocardium, which depends on a complex interaction of the intrinsic properties of individual cells and electrotonic coupling between cells.36 37 Each cardiac cell was presumed to have a unique set of electrical properties (variable upstroke velocity and APDs) on the basis of its geometry and distribution of ionic channels. Because the spatial distribution of these intrinsic cellular properties was not known, models assumed either a random two-dimensional distribution of APDs37 or a one-dimensional strand of cells with uniform APDs arranged with an abrupt boundary where cells with short APDs were coupled to cells with long APDs.36 In both cases, electrical coupling between cells was shown to modulate the intrinsic APDs of cells. Under normal conditions (low-resistance electrotonic coupling between cells), the local variation of APDs was averaged out and was thus not evident. Ischemia, which induces an uncoupling of cells,38 was modeled as an increase in axial resistance between cells. Simulations with high intercellular resistances produced significant variations in APDs.36 37 These simulations failed to predict repolarization patterns and heterogeneities in APDs because they did not include spatial heterogeneities of APDs and of intracellular coupling.
Thus, the present study obtained direct experimental evidence that repolarization spreads from cell to cell on the epicardium, whereas on the endocardium, it does not spread and patterns are irregular.
Origin(s) of Repolarization Wave Fronts
Whereas depolarization is normally initiated by the pacemaker cells of the SA node and follows well-established pathways, little is known regarding the origin(s) and spreading of repolarization. Optical repolarization maps indicated that some unidentified cells (with short APDs) repolarize first to drive the repolarization of adjacent cells, which would otherwise exhibit longer APDs. On the left and right ventricular epicardium, the first sites to repolarize were located near the apex at one third of the distance toward the base. These cells most likely represented breakthrough sites rather than the origin(s) of repolarization wave fronts. The possibility that repolarization began on the endocardium was tested by using the light guide and the photodiode array to simultaneously measure APs from the endocardium and epicardium on opposite sides of left ventricles. APDs on the endocardium were consistently shorter by 9±1.8 ms compared with the epicardium. Delays between endocardial and epicardial activation and repolarization were brief, as expected, given the wall thickness of ventricles. These delays were nevertheless significant because (1) the sampling rate of signals was sufficiently rapid (0.7 ms per frame), (2) the signal processing techniques identified unique activation and repolarization timepoints, and (3) these brief activation and repolarization delays were reproducible from beat to beat when we compared endocardial to epicardial delays or delays between adjacent diodes on the epicardium.
In perfused ventricular sheets, APDs were heterogeneous on the epicardium, being shorter near the apex than the base by 26 ms (n=8). On the endocardium, APDs were consistently shorter and more homogeneous than on the epicardium, being shorter near the apex than the base by 9 ms (n=8). These findings on perfused sheets supported the equivalent measurements on intact hearts using the light guide and photodiode array to determine APDs simultaneously on the endocardium and epicardium.
On the endocardium, the first site to depolarize was the first to repolarize, but repolarization appeared random. The random repolarization on the endocardium was attributed to the greater homogeneity of APDs. If all cells had identical APDs, one would predict that the first cells to depolarize would be the first to repolarize, and the spread of repolarization and activation would be similar. However, slight heterogeneities in APDs would produce random repolarization patterns that spread slightly slower (13.8 to 23.0 ms) compared with activation (11.5 to 15.4 ms, Fig 12A⇑ through 12D). These data suggested that repolarization in intact hearts was initiated on the endocardium at a region that repolarized first because it depolarized first. Repolarization could then spread transmurally, break through to the epicardium, and spread anisotropically along the epicardial surface (Figs 5⇑ and 6⇑).
Repolarization Patterns Measured From Dog and Guinea Pig Hearts
Repolarization has been extensively mapped in dog hearts with extracellular electrodes, but no equivalent studies exist in guinea pig hearts. Repolarization measured by ARIs has indicated that repolarization spreads passively and that once the heart is activated, repolarization duration is independent of the activation sequence.39 In isolated blood-perfused dog hearts, repolarization patterns measured from the anterior and posterior surfaces were largely independent of the stimulation protocol.39 The present measurement in guinea pig hearts demonstrates that in the epicardium there is a substantial degree of independence between activation and repolarization, as in dog hearts. A number of factors could account for the striking correlation between repolarization and fiber orientation in guinea pig hearts, which was not reported in dog hearts: (1) The present studies examined the ventricular free walls where fiber orientations are highly regular compared with the anterior and posterior septal surfaces. (2) In the dog heart, the repolarization sequence was highly dependent on the activation sequence. This was mainly due to the long activation time of the dog ventricles (30 to 60 ms), which tended to be greater than the intrinsic heterogeneities of APDs. In the smaller guinea pig heart, the repolarization sequence was independent of the activation sequence, because inhomogeneities of APDs were larger than the duration of time it took to depolarize the ventricles (≈5 ms in sinus rhythm, ≈20 ms during epicardial pacing). (3) There may be species differences in the expression and/or distribution of ionic channels. For instance, in dogs there are known heterogeneities in the distribution of Ito, resulting in greater Ito and longer APDs in epicardial than endocardial cells.40 In contrast, guinea pig myocytes lack Ito, yet studies of isolated myocytes reported longer APDs from cells isolated from the base than the apex41 and longer APDs in intramural cells compared with endocardial and epicardial cells.42 The present optical maps are consistent with heterogeneities of APDs between myocytes from the apex and the base. (4) In situ studies of dog hearts using monophasic AP recordings demonstrated an inverse relation between APDs and activation time. When activation patterns were altered by pacing the epicardium, this inverse relation was lost but returned 2 hours later. This phenomenon, called cardiac memory, indicated that under certain conditions repolarization was modified by activation after substantial time delays.6 (5) The latter studies bring up another complexity, because APDs are known to depend on the mode of contraction. As shown in papillary muscles, isotonic shortening prolongs APDs, whereas isometric contractions shorten APDs.43 Such a mechanism could produce different disparities of APDs along ventricles and from endocardium to epicardium and could contribute to the different results in dog hearts in situ compared with Langendorff-perfused guinea pig and dog (blood-perfused) hearts, which do not perform mechanical work.39
Mapping of Repolarization by Electrode and Optical Techniques
The independence of activation and repolarization could be detected by optical and ARI techniques, since both methods measure repolarization time points [optical, (d2F/dt2)max of AP downstrokes; ARI, (dV/dt)max of T waves] that are coincident with the refractory period of APs.8 25 30 In contrast with epicardial repolarization, the first site to repolarize on the endocardium was the first to activate, and APDs were more uniform on the endocardium compared with the epicardium. ARI measurements were not reported for the dog endocardium.
Optical techniques offer advantages as well as disadvantages compared with extracellular electrodes: (1) The voltage-dependent response of the dye is equivalent to an intracellular recording, dependent on the local transmembrane potential, and does not detect potential changes beyond the molecular distance of the dye. In contrast, unipolar and, to a lesser extent, bipolar recordings integrate potential changes beyond their physical dimensions. (2) Optical signals are independent of the direction of spread, whereas in bipolar recordings the shape and kinetics of the signals depend on the direction of spread. (3) The dyes’ responses are not influenced by electrical artifacts. (4) Optical maps are readily obtained at high density, with virtually no dead space between recording sites; consequently, the electrical activity of the entire surface of the epicardium is investigated with equal weight. With optical recordings, there are concerns regarding phototoxic effects and motion artifact. Some of these problems have been overcome with better dyes, chamber designs, and signal processing to measure repolarization time points; however, optical methods have not been successfully applied in situ (blood-perfused hearts) or in the clinical setting or to measure intramural APs, as can be done with plunge electrodes.
Put together, optical maps of APs provide a set of rules that guide activation and repolarization pathways in guinea pig hearts under physiological sinus rhythm or under stimulation at a number of different epicardial sites. From these guidelines, three-dimensional patterns of propagation can be predicted for activation and repolarization in intact hearts.
Selected Abbreviations and Acronyms
|αF||=||average fiber orientation|
|Ito||=||transient outward K+ current|
|Lmax||=||fiber length at maximum developed tension|
|Lo||=||fiber length at zero developed tension|
This study was supported by grants from the Whitaker Research Foundation and the American Heart Association (No. 87-1065) to Dr Salama and the Western Pennsylvania Affiliate of the American Heart Association to Dr Kanai. The authors would like to thank William Hughes for machining chambers and optical components, Gregory Szekeres and James VonHedemann for manufacturing electronic components, Gregory Shander for his assistance with the histology, William Bolish for software support, Richard Lombardi for his expertise in computer hardware and software, and Anne Crews, Dr Madison Spach, and Dr Harold Strauss for their critical review of the manuscript.
Previously published as preliminary results in abstract form (Circulation. 1988;78:II-413 and 1990;82:III-649).
- Received December 16, 1993.
- Accepted June 13, 1995.
- © 1995 American Heart Association, Inc.
Scher AM, Spach MS. Cardiac depolarization and repolarization and the electrocardiogram. In: Berne RM, ed. Handbook of Physiology, Section 2: The Cardiovascular System. Baltimore, Md: Williams & Wilkins Co; 1979:357-392.
Spach MS, Barr RC. Ventricular intramural and epicardial potential distributions during ventricular activation and repolarization in the intact dog. Circ Res. 1975;37:243-257.
Spach MS, Barr RC. Analysis of ventricular activation and repolarization from intramural and epicardial potential distributions for ectopic beats in the intact dog. Circ Res. 1975;37:830-843.
Costard-Jäckel A, Goetsch B, Antz M, Franz MR. Slow and long-lasting modulation of myocardial repolarization produced by ectopic activation in isolated rabbit hearts: evidence for cardiac ‘memory.’ Circulation. 1989;80:1412-1420.
Millar CK, Kralios FA, Lux RL. Correlation between refractory periods and activation-recovery intervals from electrograms: effects of rate and adrenergic interventions. Circulation. 1985;72:1372-1379.
Haws CW, Lux RL. Correlation between in vivo transmembrane action potential durations and activation-recovery intervals from electrograms: effects of interventions that alter repolarization time. Circulation. 1990:81;281-288.
Watanabe T, Rautaharju PM, McDonald TF. Ventricular action potentials, ventricular extracellular potentials, and the ECG of guinea pig. Circ Res. 1985;57:362-373.
Spaggiari MS, Bariffi S, Macchi E, Traversa M, Arisi G, Taccardi B. The effects of cardiac electric anisotropy on epicardial potential fields during ventricular repolarization. Jpn Heart J. 1986;27:217-223.
Steinhaus BM. Estimating cardiac transmembrane activation and recovery times from unipolar and bipolar extracellular electrograms: a simulation study. Circ Res. 1989;64:449-462.
Kuo CS, Munakata K, Reddy CP, Surawicz B. Characteristics and possible mechanism of ventricular arrhythmia dependent on the dispersion of action potential durations. Circulation. 1983;67:1356-1367.
Salama G, Lombardi R, Elson J. Optical maps of action potential propagation and NADH fluorescence in intact working hearts. Am J Physiol. 1987;252:H384-H394.
Salama G. Optical measurements of transmembrane potential in heart. In: Loew L, ed. Spectroscopic Probes of Membrane Potential. New York, NY: CRS Uniscience; 1988;3:137-199.
Salama G, Morad M. Merocyanine 540 as an optical probe of transmembrane electrical activity in the heart. Science. 1976;191;485-487.
Li T, Sperelakis N, Teneick RE, Solaro JR. Effects of diacetyl monoxime on cardiac excitation-contraction coupling. J Pharmacol Exp Ther. 1985;232:688-695.
Davidenko JM, Kent PF, Dante RC, Michaels DC, Jalife J. Sustained vortex-like waves in normal isolated ventricular muscle. Biophysics. 1990;87:8785-8789.
Dillon SM. Optical recording in the rabbit heart shows that defibrillation strength shocks prolong the duration of depolarization and the refractory period. Circulation. 1991;80:215-216.
Lab MJ. Mechanically dependent changes in action potentials recorded from the intact frog ventricle. Circ Res. 1978;42:519-528.
Lab MJ. Contraction-excitation feedback in myocardium: physiological basis and clinical significance. Circ Res. 1982;50:757-766.
Salama G, Rosenbaum D, Kanai A, Cohen RM, Kaplan DT. Data analysis techniques for measuring spatial inhomogeneity in repolarization using optical transmembrane potentials. IEEE Eng Med Biol Soc. 1989;1:222-223.
Kanai A. Patterns of Electrical Activation and Repolarization in Mammalian Heart, Measured With Voltage Sensitive Dyes and Imaging Techniques. Pittsburgh, Pa: University of Pittsburgh School of Medicine; 1992. Thesis.
Salama G, Kanai A, Efimov IR. Subthreshold stimulation of Purkinje fibers interrupts ventricular tachycardia in intact hearts: experimental study with voltage-sensitive dyes and imaging techniques. Circ Res. 1994;74:604-619.
Kanai A, Salama G. Maps of optical action potentials and fiber structure in guinea pig ventricles. Circulation. 1988;78(suppl II):II-413. Abstract.
Salama G, Kanai A. Voltage-sensitive dyes and optical maps of guinea pig ventricles reveal that repolarization is an active process guided by fiber orientation. Circulation. 1990;82(suppl III):III-649. Abstract.
Efimov IR, Haung DT, Rendt JM, Salama G. Optical mapping of repolarization and refractoriness from intact heart. Circulation. 1994;90:1469-1480.
Luna LG. Manual of Histological Staining Methods of the Armed Forces Institute of Pathology. 3rd ed. New York, NY: Blakiston Division, McGraw Hill; 1968:36-38.
Streeter DD, Spotnitz HM, Patel DP, Ross J Jr, Sonnenblick EH. Fiber orientation in the canine left ventricle during diastole and systole. Circ Res. 1969;24:339-347.
Carew TE, Covell JW. Fiber orientation in hypertrophied canine left ventricle. Am J Physiol. 1979;263:H487-H493.
Frazier DW, Krassowska W, Chen PS, Wolf PD, Danieley ND, Smith WM, Ideker RE. Transmural activation and stimulus potentials in three-dimensional anisotropy canine myocardium. Circ Res. 1988;63:135-146.
Nielsen PNF, Le Crice IJ, Smaill BH, Hunter PJ. Mathematical model of geometry and fibrous structure of the heart. Am J Physiol. 1991;260:H1365-H1378.
Joyner RW. Modulation of repolarization by electrotonic interactions. Jpn Heart J. 1986;27:167-183.
Lesh MD, Pring M, Spear JF. Cellular uncoupling can unmask dispersion of action potential duration in ventricular myocardium: a computer modeling study. Circ Res. 1989;65:1426-1440.
Kléber AG, Riegger CB, Janse MJ. Electrical uncoupling and increase of extracellular resistance after induction of ischemia in isolated, arterially perfused rabbit papillary muscle. Circ Res. 1987;61:271-279.
Burgess MJ, Baruffi S, Spaggiari S, Macchi E, Lux R, Taccardi B. Determination of activation and recovery sequences and local repolarization durations from distant electrocardiographic leads. Jpn Heart J. 1986;27(suppl):205-216.
Livotsky S, Anzelevitch C. Transient outward current prominent in canine ventricular epicardium but not endocardium. Circ Res. 1988;62:116-126.
Watanabe T, Delbridge LM, Bustamante JO, McDonald TF. Heterogeneity of the action potential in isolated rat ventricular myocytes and tissue. Circ Res. 1983;52:280-290.