Simultaneous maps of optical action potentials and calcium transients in guinea-pig hearts: mechanisms underlying concordant alternans

Abstract

1. The mechanisms underlying electro-mechanical alternans caused by faster heart rates were investigated in perfused guinea-pig hearts stained with RH237 and Rhod-2 AM to simultaneously map optical action potentials (APs) and intracellular free Ca2+ (Ca2+i). 2. Fluorescence images of the heart were focused on two 16 x 16 photodiode arrays to map Ca2+i (emission wavelength (lamdda;em) = 585 +/- 20 nm) and APs (lamdda;em > 715 nm) from 252 sites. Spatial resolution was 0.8 mm x 0.8 mm per diode and temporal resolution 4000 frames s-1. 3. The mean time-to-peak for APs and [Ca2+]i was spatially homogeneous (8.8 +/- 0.5 and 25.6 +/- 5.0 ms, respectively; n = 6). The durations of APs (APDs) and Ca2+i transients were shorter at the apex and progressively longer towards the base, indicating a gradient of ventricular relaxation. 4. Restitution kinetics revealed increasingly longer delays between AP and Ca2+i upstrokes (9.5 +/- 0.4 to 11.3 +/- 0.4 ms) with increasingly shorter S1-S2 intervals, particularly at the base, despite nearly normal peak [Ca2+]i. 5. Alternans of APs and Ca2+i transients were induced by a decrease++ in cycle length (CL), if the shorter CL captured at the pacing site and was shorter than refractory periods (RPs) near the base, creating heterogeneities of conduction velocity. 6. Rate-induced alternans in normoxic hearts were concordant (long APD with large [Ca2+]i) across the epicardium, with a magnitude (difference between odd-even signals) that varied with the local RP. Alternans were initiated by gradients of RP, producing alternans of conduction that terminated spontaneously without progressing to fibrillation.

Figure 1. Optical apparatus and emission spectra of Rhod-2 and RH237

A, schematic diagram of optical apparatus. Light from two 100 W tungsten-halogen lamps was collimated, passed through 520 ± 20 nm interference filters, and focused on the heart. Fluorescence from the stained heart was collected by a camera lens and passed through a dichroic mirror to split the emission wavelength below and above 630 nm. Wavelengths below 645 nm were passed through a 585 ± 20 nm interference filter and those above through a 715 nm cut-off filter, and the two images of the heart were focused on two 16 × 16 photodiode arrays. hv, excitation light. B, emission spectra of Rhod-2 and RH237 from the stained heart. The Rhod-2 emission spectrum has a peak at 585 nm and returns to baseline at 660 nm. The RH237 emission spectrum is broad but the voltage-dependent spectral change during APs (i.e. the action spectrum) was a decrease in fluorescence from 670 to 760 nm, a longer wavelength range than that of most potentiometric dyes. Based on these spectra, filter sets were chosen to eliminate cross-talk between the two dyes without compromising the signal-to-noise ratio of the fluorescence signals.

Figure 2. Symbolic maps of APs and Ca²⁺_i recorded from the anterior epicardium

A, diagram showing the anterior surface of the guinea-pig heart and the field-of-view of the array used throughout the study. The top edge of the array is aligned with the base of the heart and the bottom edge with the apex. RA, right atrium; RV, right ventricle; LA, left atrium; LV, left ventricle. B, superposition of AP and Ca²⁺_i traces recorded at 300 ms CL from a voltage and a Ca²⁺_i diode located at the centre of the two arrays. For each voltage diode, there is an equivalent Ca²⁺_i diode and fewer than 5 % of the sites detected contraction artifacts too large to permit a quantitative analysis of APDs and [Ca²⁺]_i,75. C and D, symbolic maps of two photodiode arrays. Each square box represents a diode on the array and with APs and Ca²⁺_i transients drawn in their respective location on the voltage and Ca²⁺_i arrays. These APs and Ca²⁺_i transients were measured simultaneously from 252 sites and are shown as they appear on the computer monitor.

Figure 6. Response characteristics of APs and Ca²⁺_i transients during a premature impulse

A, simultaneous recordings of APs and Ca²⁺_i transients from two diodes in register to measure voltage and Ca²⁺_i from the same site on the heart. The heart was paced for 20 beats at a basic CL, with S1-S1 = 300 ms, followed by a single premature impulse at variable S1-S2 intervals. B, time delay between APs and Ca²⁺_i as a function of the S1-S2 interval. The spatio-temporal delays between the AP upstroke and the rise in Ca²⁺_i were analysed by plotting the delay between (dF/dt)_max of the voltage and Ca²⁺_i traces as a function of the S1-S2 interval for 6 sites at the apex and 6 sites at the base. There were no significant differences in AP to Ca²⁺_i delays at different sites on the heart, except for the shortest S1-S2 intervals that could capture an AP that spread on the heart. However, at the shortest S1-S2 intervals (182 ms), delays between APs and Ca²⁺_i were significantly longer at the base than at the apex (P < 0.01), perhaps because the premature pulse impinges more on the refractory period of the tissue at the base than at the apex. Each data point represents the mean delay from 6 sites (or diodes) ±s.d.; the plot was reproduced in 4 separate hearts. ANOVA analysis and Student’s t test with Bonferroni’s correction yielded statistically significant values (*P < 0.01). C, comparison of AP and Ca²⁺_i at the base and apex at the shortest S1-S2 interval. AP upstroke (continuous lines) and Ca²⁺_i rise (dashed lines) are shown for a site at the base (top traces) and at the apex (bottom traces) elicited by the shortest S1-S2 interval in the plot shown in B. Note the marked slowing down of the rise in Ca²⁺_i at the base compared with the apex.

Figure 7. Restitution kinetics of APs and Ca²⁺_i at the base and apex of the heart

A and B, duration of APs (A) and Ca²⁺_i transients (B) plotted as a function of the S1-S2 interval for sites at the base and the apex of the heart. The durations of the voltage and Ca²⁺_i traces decreased with decreasing S1-S2 intervals and were longer at the base than at the apex, except for the shortest S1-S2 interval, because of the longer refractory period at the base than the apex. C-F, similarly, the amplitudes and rates-of-rise of APs (C and E) and Ca²⁺_i transients (D and F) decreased with decreasing S1-S2 intervals but the effects were more pronounced at the base than at the apex. Each point represents the mean value derived from 6 diodes at the base and 6 diodes at the apex; error bars represent s.d. ANOVA analysis and Student’s t test with Bonferroni’s correction yielded statistically significant values; *P < 0.01.

Figure 4. Calibration of [Ca²⁺]_i

A, minimum (F_min) and maximum (F_max) fluorescence levels from Rhod-2 trapped in the heart were measured as described in Methods and plotted on a relative intensity scale with a Ca²⁺_i transient (F). B, as described in Methods, a Ca²⁺_i transient was calibrated in terms of free cytosolic Ca²⁺, based on F_min, F_max and K_d.

Figure 3. Lack of cross-talk between optical measurements of APs and Ca²⁺_i

A, signals recorded from 2 diodes in register to detect APs and Ca²⁺_i when the heart was stained with RH237 but not Rhod-2. B, as for A but for a heart stained with Rhod-2, not RH237. Note the lack of cross-talk between the voltage and Ca²⁺_i diode. C and D, effects of ryanodine on AP and Ca²⁺_i recordings. To verify that the Rhod-2 signal is a measure of [Ca²⁺] released from RyRs, ryanodine (10 μm) was added to the perfusate to test the selectivity of its actions on Ca²⁺_ivs. the AP. C, control recordings of APs and Ca²⁺_i transients from diodes in exact register. D, recordings from the same two diodes after perfusion with ryanodine. Note that, after ryanodine treatment, Ca²⁺_i transients became rounded in shape, lost the rapid rise in [Ca²⁺]_i immediately after the AP upstroke and that peak [Ca²⁺]_i was markedly reduced to 27 ± 3 % (n = 3) of the original signal in 10 min. In contrast, APs were essentially unchanged, except for a prolongation of APD caused by the decrease in [Ca²⁺]_i.

Figure 5. Simultaneous maps of activation, repolarization and duration for APs and Ca²⁺_i transients

Isochronal maps depicting the direction and rate of propagation of voltage and Ca²⁺_i, as described in Methods. Each map is from the anterior epicardium, as in Fig. 2A and represents analysis of the 252 APs (A-D) and the equivalent Ca²⁺_i transients (A′-D’) recorded simultaneously from the same heart. The light to dark shading shows the progression from earliest activation/repolarization and from short to long duration. A and A’, activation maps for a guinea-pig heart under sinus rhythm. Activation breaks through over a large region of epicardium, originating from the specialized conduction system on the endocardium leading to rapid depolarization within ≈3 ms. Ca²⁺_i elevation follows the same pattern and time course. B and B’, activation maps drawn for the same heart when stimulated at 300 ms from the left ventricle (right of the field of view). APs propagate from the stimulus site in ≈16 ms according to the orientation of epicardial fibre. Ca²⁺_i activation follows AP activation with a time delay of 10 ms in this heart. C and C’, maps of AP repolarization and of Ca²⁺_i recovery ([Ca²⁺]_i,75). Repolarization begins at the apex on the left ventricle 163 ms after the first activation and spreads towards the base or outflow track on the right ventricle. Ca²⁺_i recovery follows a similar pattern with slower recovery times. Similar maps to those shown in C and C’ were recorded when the heart was under sinus rhythm or paced on the right or left ventricle, indicating that repolarization is independent of activation and is driven by spatial heterogeneities of K⁺ channels responsible for the repolarization of the AP. D and D’, maps of APDs and durations of Ca²⁺_i transients ([Ca²⁺]_i,75). APDs are shorter at the apex than at the base and the recovery and duration of Ca²⁺_i were similar to the electrical repolarization pattern and reveal a gradient of relaxation from apex to base. Isochrones are 1 ms apart for A, B, A’ and B’ and 2 ms apart for C, C’, D and D’.

Figure 8. Optical APs and Ca²⁺_i transients during a change in rate (long to short CL)

A, traces illustrating the changes in APs and Ca²⁺_i transients that occur during a rate change, by pacing the epicardium at long CL (300 ms) for 20 beats (last 2 beats shown) followed by 10 beats at short CLs (167 ms). Note that the resting potential (top trace) and diastolic Ca²⁺_i are elevated at short CLs and both oscillate between even and odd beats when the rate was increased. B and C, APs and Ca²⁺_i transients, respectively, and their first derivatives shown on an expanded time scale. An increase in rate always produced concordant alternans since the larger Ca²⁺_i transient is always associated with the longer AP and the larger first derivative of Ca²⁺_i is always associated with the larger first derivative of the AP upstroke. Such concordant alternans could be elicited repeatedly by shifting from long to short CLs with the same heart (8-10 times) and the same findings were obtained in 6 separate hearts.

Figure 9. AP and Ca²⁺_i alternans at the apex and base

The rates-of-rise of APs and Ca²⁺_i transients were normalized with respect to the control basic beats and then plotted as a function of odd and even beats during concordant alternans elicited by a rate change from 300 to 167 ms CL, as in Fig. 8. Each data point on the plot represents the average rate-of-rise from 8 diodes recording from the base or apex and was plotted against the beat number following the change in rate. Note that the amplitudes of AP and Ca²⁺_i alternans are markedly larger at the base of the heart (A and B) compared with the apex (C and D). Moreover, there was a more rapid damping of alternans at the apex than at the base. In 6 out of 6 hearts, the amplitude of alternans dampened and self-terminated within 20 beats. ANOVA analysis between odd and even beats showed statistically significant differences at the base (P < 0.01) but not at the apex of the heart.

Figure 10. Distribution of the magnitude of alternans and its correlation with APD

The rates-of-rise of APs and Ca²⁺_i transients during alternans were normalized with respect to the control basic beats (as in Fig. 9). The magnitude of alternans for voltage (ΔAP) and Ca²⁺_i (ΔCa²⁺_i) was calculated for each site (see Methods) and mapped in A and B, respectively. Contour lines of the magnitude of alternans are drawn for every 5 % difference in the normalized rate-of-rise. White to black represents increasingly larger magnitude of alternans. C and D, plots of the magnitude of alternans versus the APD at each site for voltage (C) and Ca²⁺_i (D). The correlation coefficients are 0.79 (P < 0.01) and 0.70 (P < 0.01) for ΔAP and ΔCa²⁺_i, respectively, indicating a high degree of correlation between the magnitude of alternans and the local refractory period of the heart.

Figure 11. The first rapid beats encounter lines of functional conduction block that initiate alternans

Maps of activation and local velocity vectors are displayed for the last basic beat (top 4 panels; CL = 300 ms) and the first beat at short CL (bottom 4 panels; CL = 167 ms) when the heart was paced from the left (A) or the right (B) side. During ‘control’ beats (top panels), APs propagated from the stimulus site (square pulse) in 15-17 ms. Local velocity vectors are uniformly oriented away from the stimulus site and reveal no rapid changes in direction between adjacent sites. On the other hand, the first rapid stimulus, captured at the site of stimulation, propagated and then encountered a functional line of conduction block (bottom panels). Isochronal lines and local velocity vectors show abrupt changes, respectively, in density and orientation at the line of block. Wave propagation slowed down at the base resulting in curved wavefronts around the arc of conduction block. Isochrones are 1 ms apart and all local velocity vectors were normalized to focus on the changes in orientation rather than absolute values. In 6 out of 6 hearts, the first rapid beat that captured at short CL propagated around an arc of functional conduction block, as illustrated in the bottom panels.

Figure 12. Concordant alternans are linked to alternations of conduction velocity

Activation patterns are displayed as isochronal maps and maps of local velocity vectors for odd and even beats of the alternans shown in Figs 8 and 9. Isochronal maps (left) and local velocity vectors (right) compare the propagation of APs for even (top panels) and odd (bottom panels) beats. Activation patterns alternated between even and odd beats because the conduction velocity was consistently and markedly slower at the base (28.8 ± 8.0 cm s⁻¹) than at the apex (66.9 ± 8.3 cm s⁻¹) of the heart. Isochronal lines are 2 ms apart.