Basal Spontaneous Firing of Rabbit Sinoatrial Node Cells is Regulated by Dual Phosphodiesterase 3 and 4 Activation

Abstract

Background

Spontaneous firing of sinoatrial node cells (SANC) is regulated by cAMP-mediated, PKA-dependent (cAMP/PKA) local subsarcolemmal Ca²⁺ releases (LCRs) from ryanodine receptors (RyR). LCRs occur during diastolic depolarization (DD) and activate an inward Na⁺/Ca²⁺ exchange current that accelerates DD rate prompting the next action potential (AP). Phosphodiesterases (PDEs) regulate cAMP-mediated signaling; PDE3/PDE4 represent major PDE activities in SANC, but how they modulate LCRs and basal spontaneous SANC firing remains unknown.

Methods and Results

PDE3A, PDE4B and PDE4D were the major PDE subtypes (real-time PCR, western blot) expressed in rabbit SANC, and PDE3A (immunostaining) was co-localized with α-actinin, PDE4D, SERCA and phospholamban in Z-lines. Inhibition of PDE3 (cilostamide) or PDE4 (rolipram) alone increased spontaneous SANC firing (perforated patch) by ∼20% (P<0.05) and ∼5% (P>0.05) respectively, but concurrent PDE3+PDE4 inhibition increased spontaneous firing by ∼45% (P<0.01), indicating synergistic effect. Inhibition of PDE3 or PDE4 alone increased L-type Ca²⁺ current (I_Ca,L) by ∼60% (P<0.01) or ∼5% (P>0.05), respectively, and phospholamban phosphorylation by ∼20% (P>0.05) each, but dual PDE3+PDE4 inhibition increased I_Ca,L by ∼100% (P<0.01) and phospholamban phosphorylation by ∼110%(P<0.05). Dual PDE3+PDE4 inhibition increased LCR number and size (confocal microscopy; P<0.01), reduced SR Ca²⁺ refilling time (P<0.01) and the LCR period (time from AP-induced Ca²⁺transient to subsequent LCR; P<0.01), leading to decrease in spontaneous SANC cycle length (P<0.01). When RyR were disabled by ryanodine and LCRs ceased, dual PDE3+PDE4 inhibition failed to increase spontaneous SANC firing.

Conclusions

Basal cardiac pacemaker function is regulated by concurrent PDE3+PDE4 activation which operates in a synergistic manner via decrease in cAMP/PKA phosphorylation, suppression of LCR parameters, prolongation of the LCR period and spontaneous SANC cycle length.

Graphical abstract

Introduction

Normal automaticity of the heart is initiated within cardiac pacemaker, the sinoatrial (SA) node; excitation then propagates to atria and ventricles to trigger cardiac muscle contraction, which delivers blood to the body. Spontaneous beating of the SA node is emanated from beating of SA node pacemaker cells (SANC), which spontaneously generate action potentials (AP) due to gradual depolarization of the membrane potential during diastole, i.e. diastolic depolarization (DD).¹ Spontaneous firing of SANC is critically dependent on surface membrane ion channels and sarcoplasmic reticulum (SR) generated local subsarcolemmal Ca²⁺ releases (LCR). Rhythmic LCRs appear during late DD and activate an inward Na⁺/Ca²⁺ exchange current (I_NCX), which accelerates DD rate and prompts the generation of subsequent AP.² The ionic currents in SANC include hyperpolarization activated “funny” current I_f, L-type and T-type Ca²⁺ currents (I_Ca,L, I_Ca,T), delayed rectifier potassium current (I_K), Na⁺-Ca²⁺ exchange current (I_NCX), etc. Both ionic channels and intracellular SR Ca²⁺ cycling in SANC work together to guarantee stability and flexibility of cardiac pacemaker function.³

cAMP is a ubiquitous second messenger that modulates substantial number of cell processes, e.g. cAMP-mediated activation of PKA-dependent phosphorylation of multiple proteins. Constitutive activation of adenylyl cyclases (ACs) in rabbit SANC generates high basal level of both cAMP and cAMP-mediated PKA-dependent phosphorylation, which are required for generation of spontaneous LCRs and normal spontaneous beating of SANC.^4,5 Although high basal cAMP production in SANC might indicate low cAMP degradation by phosphodiesterases (PDE), an increase in cAMP level and spontaneous SANC beating rate after suppression of basal PDE activation by broad-spectrum PDE inhibitor IBMX exceeds that in response to stimulation of β-adrenergic receptors (β-AR) with isoproterenol. This indicates the presence of high basal PDE activity in SANC.⁵

More than 60 PDE isoforms, that comprise 11 families (PDE1-11), exist in mammalian cells, and at least four families PDE1-PDE4 can hydrolyze cAMP in the heart. PDE1 is activated by Ca²⁺/calmodulin, PDE2 is stimulated by cGMP, PDE3 is inhibited by cGMP and PDE4 is specific for cAMP. Although PDE3 can hydrolyze both cAMP and cGMP, the catalytic rates for cAMP are 5-10-fold higher, than for cGMP, which makes PDE3 highly specific for cAMP.⁶

Inhibition of PDE3 causes sinoatrial tachycardia in guinea pigs,⁷ rabbits,^5,8 dogs⁹ and humans.¹⁰ PDE4 is the dominant PDE isoform in the murine heart,⁶ and inhibition of either PDE3 or PDE4 produces sinoatrial tachycardia in mice¹¹ and rats.⁸ Several ionic currents involved in the generation of the DD are regulated by PDEs, i.e. inhibition of PDE3 in rabbit SANC increases I_Ca,L, I_K and shifts voltage dependence of I_f activation to more positive potentials.^5,12,13 Funny current I_f is directly activated by cAMP mostly through HCN4 channel.¹⁴ LCRs are also regulated by PDEs, i.e. PDE inhibition reduces the LCR period, shifting LCR occurrence to earlier times during DD, and increases LCR number and size as RyR activation becomes more synchronized via RyR recruitment. The earlier and stronger LCR-generated Ca²⁺ release results in an increase and earlier activation of I_NCX, acceleration of the DD rate and increase in the spontaneous SANC beating rate.⁵

There is a growing evidence to suggest that while individual PDE3 or PDE4 inhibition have minor or no effect on their own, combined PDE3+PDE4 inhibition could produce a large synergistic response, creating effect which is greater than the simple sum of separate PDE3 and PDE4 inhibition.^15,16 Synergistic effects of concurrent PDE3+PDE4 inhibition have been previously observed in variety of cell types, including glucose uptake by brown adipose tissue,¹⁶ regulation of smooth muscle cell motility¹⁷ and increase in contractility by rat VM¹⁸ or right atrium.¹⁹

PDE3 and PDE4 represent major PDE activities in the rabbit SA node, i.e. their combined activity in cytosolic or SR fraction accounts for ∼50% and ∼90% of total cAMP-PDE activity, respectively, while contribution from other PDE subtypes is relatively small.²⁰ However, how PDE3 and PDE4 regulate spontaneous beating rate of cardiac pacemaker cells and whether there is synergistic effect of concurrent PDE3 and PDE4 activation remains unknown. The aims of the present study were to determine: (1) the major PDE (i.e. PDE3 and PDE4) subtypes expressed in rabbit SANC; (2) how major PDE3 and PDE4 subtypes are distributed within SANC; (3) whether PDE3 and PDE4 work in a synergistic manner to regulate spontaneous SANC firing; and, if so, (4) what specific targets are modulated by concurrent PDE3+PDE4 activation.

Materials and Methods

An extensive description of ‘Materials and Methods’ is provided in the online Data Supplement. All experiments involving rabbit SANC or tissue samples were carried out in accordance with the NIH Guide for the Care and Use of Laboratory Animals. Selected methods including isolation of rabbit SANC and measurements of LCR characteristics were published before and are available online at PubMed: https://www.ncbi.nlm.nih.gov/pubmed/?term=vinogradova+TM+circ+rec+2002+90%3A73 https://www.ncbi.nlm.nih.gov/pubmed/?term=vinogradova+TM+circ+rec+2008 Other data, analytic methods, and study materials will be made available to other researchers on request for the purpose of reproducing the results or replicating the procedure.

SA node cell preparations and electrophysiological recordings

SANC or VM were isolated from rabbit hearts, as previously described;⁵ only regular spontaneously beating SANC were selected for experiments. SANC were continuously superfused with the Tyrode solution, which was switched to the Tyrode solution containing chemicals employed in the experiment. The bath temperature was maintained at 35±0.5°C. The perforated patch-clamp technique and whole cell patch-clamp technique were employed to record APs and I_Ca,L, respectively. Spontaneous beating of SANC was recorded for at least 10-15 minutes in the basal state, followed by drug application. Effects of PDE inhibitors on the beating rate were also studied in the presence of 3 μmol/L ryanodine (Ry) or I_f current inhibitors 5 μmol/L ivabradine or 2 mmol/L Cs⁺.

Confocal Ca²⁺ transients and LCRs

To measure cytosolic Ca²⁺ SANC were loaded with fluo-3 AM (Molecular Probes, Eugene, OR); images were recorded using confocal microscopy in the line-scan mode, as previously described.²¹

Cell permeabilization

A subset of SANC were permeabilized with 0.01% saponin, LCR number was normalized per 100 μm of the linescan image and 1 second time interval, as previously described.²²

RNA extraction and RT-QPCR

RNA was extracted from isolated rabbit SANC (n=4-9 pooled samples, each sample collected from 3 rabbits) or left ventricular myocytes (VM) (n=6 rabbits) with RNeasy Mini Kit (Qiagen, Valencia, CA) and DNAse on column digestion. For cDNA preparation we used MMLV reverse transcriptase (Promega, WI, USA). VM and pooled SANC samples were tested via RT-QPCR, using ABI Prism 7900HT Sequence Detection System (Applied Biosystems, USA).

Western blotting

To detect different PDE subtypes SA node and left ventricular (LV) tissues were frozen in liquid nitrogen, powdered and resuspended in RIPA lysis buffer. PDE3A, PDE4A, PDE4B and PDE4D in the SA node and LV were detected using PDE3A, PDE4A and PDE4D antibody (Abcam, Cambridge, MA) and custom made PDE4B antibody (GenScript, Piscataway, NJ). The detection of PLB phosphorylation at the PKA-dependent site was performed in isolated SANC, using a phosphorylation specific P-Ser¹⁶ PLB antibody (Badrilla, UK) as previously described.⁴

Immunostaining

Intact rabbit SANC were incubated with primary anti-PDE3A antibody together with either anti-α-actinin (Sigma-Aldrich, USA) or anti-PDE4B or anti-PDE4D or anti-SERCA2 (ThermoFisher, USA) or anti-PLB antibodies (Badrilla, UK). Dual confocal images of central sections of SANC were obtained via Zeiss LSM 510 (Carl Zeiss Inc., Germany). Images were processed via LSM 5 Image Browser (Carl Zeiss Inc., Germany) and intensity of immunofluorescence was plotted using ImageJ software (1.8V, Wayne Rasband, NIH).

Statistical analysis

Data were presented as mean±SEM. The statistical significance of effects was evaluated with PRISM software, i.e. Student t-test, column statistics or analysis of variance (ANOVA) where appropriate. A value of P<0.05 was considered statistically significant.

Results

RNA abundance, protein expression and distribution of different PDE subtypes in rabbit SANC

Total RNA from SANC and VM was reverse transcribed to generate complementary DNA (cDNA), and relative abundance of cDNA from different PDE transcripts was measured with RT-QPCR. Although PDE1A RNA was more abundant in SANC than in VM,²³ PDE3A, PDE4B and PDE4D were the major PDE subtypes expressed in both rabbit SANC and VM (Fig.1A). There was comparable expression of PDE3A and PDE4B in rabbit SANC which surpassed expression of other PDE subtypes (Fig.1A). Consistent with the RT-QPCR data, western blots demonstrated more abundant PDE3A and PDE4A protein expression in the rabbit left ventricle, compared to the SA node (Fig.1B, Fig.1C). Expression of PDE4B protein was similar in both tissues (Fig.1D), while expression of PDE4D protein was more abundant in the rabbit SA node (Fig.1F).

Figure 1. Expression of cAMP-degrading PDEs at RNA and protein levels in cardiac pacemaker and ventricular myocardium.

A, relative expression of PDE-coding transcripts (Mean±SEM) in rabbit SANC and VM (n=4-9); one-way ANOVA with Tukey's posthoc test, adjusted *P<0.05 (SANC vs VM for each PDE subtype); +P<0.05 (PDE subtypes in SANC vs PDE3A or PDE4B); B,C,D,F, western blots of major PDE subtypes in the rabbit SA node (SAN) and left ventricle (LV): representative blots and average data (n=8) of PDE3A, PDE4A, PDE4B and PDE4D. Column statistics *P<0.05.

The intracellular distributions of the most abandunt isoforms of PDE3 (PDE3A), and PDE4 (PDE4B and PDE4D) in rabbit SANC were further examined using immunostaining. Figure 2 shows that PDE3A was detected both beneath sarcolemma and in a striated pattern within Z-lines of rabbit SANC, colocalized with the Z-line associated protein α-actinin (Fig.2A). Intensity plots in the right panel demonstrated co-distribution of PDE3A and α-actinin with an interval of 1.80±0.01μm (n=13 SANC). Co-staining of PDE3A with anti-PDE4B antibody showed higher labeling intensity in the overlay images beneath sarcolemma of SANC (Fig.2B), while PDE4D co-localized with PDE3A in striated patterns inside SANC (Fig.2C). Since PDE3A in human myocardium co-localizes with SR Ca²⁺ ATP-ase (SERCA) and phospholamban (PLB),²⁴ we studied whether the similar co-localization of PDE3A existed in rabbit SANC. Co-staining of PDE3A with SERCA or PLB antibodies showed that, indeed, PDE3A co-localized with SERCA and PLB in SANC (Fig.2D-2E). Considering that PDE4D co-localized with PDE3A (Fig.2C), which co-localized with SERCA and PLB, PDE4D should be also in the proximity of major SR proteins.

Figure 2. Distribution of major PDE3 and PDE4 subtypes in single SANC.

From the left: first column, green fluorescence staining for PDE3A; second column, red fluorescence staining for marker proteins: α-actinin (A), PDE4B (B), PDE4D (C), SERCA2 (D) and PLB (E) respectively; third column merged images of PDE3A and marker proteins in panels (A)-(E). Insets in overlay show magnification of the rectangular areas in panels (A)-(E). Fourth column: intensity plots calculated along white lines in dashed rectangles. Superimposed images (Overlay) and intensity plots showed overlapping distribution of PDE3A with α-actinin, PDE4D, SERCA and PLB along Z-lines and with PDE4B beneath sarcolemma. (F) negative control had negligible fluorescence.

Basal spontaneous firing of rabbit SANC is regulated by dual (PDE3+PDE4) activation

Though our previous report identified PDE3 as a major PDE subtype regulating basal spontaneous beating of SANC,⁵ the relatively high concentration (50μmol/L) of milrinone employed could suppress not only PDE3 activity, but also that of PDE4.^20,25 To clarify whether activation of PDE3 alone or combined PDE3+PDE4 activation regulated basal spontaneous SANC firing, we used milrinone concentration of 10 μmol/L, considered to be specific for PDE3. Milrinone at that concentration increased spontaneous SANC firing by only ∼20% (data not shown), similar to selective PDE3 inhibitor cilostamide (0.3 μmol/L),^20,26 which increased DD rate and the spontaneous SANC beating rate by ∼30% (P<0.05) and ∼20% (P<0.05), respectively (Fig.3). An acceleration of spontaneous SANC firing by selective PDE4 inhibitor rolipram (2 μmol/L),^27,28 did not reach statistical significance (Fig.3). To verify the efficacy of PDE4 inhibition by rolipram its concentration was further increased to 20μmol/L and 100μmol/L, but no further increase in the spontaneous SANC beating rate was observed (data not shown). Concurrent dual inhibition of PDE3+PDE4 by a combination of cilostamide and rolipram, however, markedly increased DD rate and spontaneous SANC beating rate by ∼70% (P<0.05) and ∼48% (P<0.01), respectively, an effect matching that of broad-spectrum PDE inhibitor IBMX (Fig.3). An acceleration of spontaneous SANC firing by concurrent dual PDE3+PDE4 inhibition by ∼2-fold exceeded the sum of increases in spontaneous firing produced by inhibition of PDE3 (∼20%) and PDE4 (∼5%) alone, indicating that dual PDE3+PDE4 activation regulated basal firing of cardiac pacemaker cells in a synergistic manner.

Figure 3. Synergistic effect of dual PDE3+PDE4 inhibition on spontaneous SANC beating rate.

A, Representative recordings of APs prior to and during inhibition of PDE3 or PDE4 alone or dual PDE3+PDE4 inhibition. B, C, Cilostamide (0.3 μmol/L) increased DD rate (from 57.7±4.2 to 73.9±6.2 mV/sec) and spontaneous beating rate (from 147.5±13.3 to 174.9±13.2 b/min, n=8 SANC, 4 rabbits), while effects of rolipram (2 μmol/L) on DD rate (from 59.0±19.0 to 62.6±19.0 mV/sec) or spontaneous firing (from 161.9±12.6 to 169.9±10.4b/min; n=7 SANC, 3 rabbits) did not reach statistical significance; combination of (cilostamide+rolipram) increased DD rate (from 52.8±4.2 to 88.8±4.3mV/sec) and beating rate (from 134.1±8.4 to 196.6±11.3 b/min, n=9 SANC, 4 rabbits); IBMX increased DD rate (from 52.5±6.8 to 88.3±8.5mV/sec) and beating rate (from 149.7±8.2 to 207.8±9.4 b/min, n=10 SANC, 6 rabbits). One-way ANOVA with Newman-Keuls's posthoc test ⁺P<0.01, *P<0.05 vs cilostomide or rolipram alone.

Basal PDE3+PDE4 activation suppressed local Ca²⁺ releases in intact SANC

During spontaneous beating AP-induced Ca²⁺ transient partially depletes SR and abolishes LCRs; when SR Ca²⁺ content is replenished by SERCA spontaneous LCRs begin to occur.²¹ The LCR period is a time-interval between prior AP-induced global Ca²⁺ transient and the LCR occurrence, it defines the time of I_NCX activation and thus generation of the next AP. Considering the essential role of LCRs for regulation of basal spontaneous SANC firing, we examined how inhibition of PDE3 or PDE4 alone or dual PDE3+PDE4 inhibition affected the LCR period and characteristics. Cilostamide markedly increased the LCR size and number per each spontaneous cycle by ∼20% each (P<0.05) and decreased the LCR period by ∼15% (P<0.05) (SMFig.1), while changes in these parameters by rolipram were minor (SMFig. 2). Concurrent inhibition of PDE3+PDE4, however, increased both the LCR size and number per each spontaneous cycle by ∼45% (P<0.01) each and decreased the LCR period by ∼40% (P<0.01) that was highly correlated with concomitant decrease in the spontaneous cycle length (Fig.4). Therefore, basal LCRs could be a primary target of dual PDE3+PDE4 regulation. To verify role of LCRs for (PDE3+PDE4)-dependent regulation of spontaneous SANC firing we used ryanodine, which locks RyRs in a sub-conductance open state, depleting the SR Ca²⁺ content and eventually eliminating LCRs. When RyR Ca²⁺ release was inhibited by ryanodine combined PDE3+PDE4 inhibition produced only a minor increase in the spontaneous SANC beating rate (Fig.5C), indicating a critical role of LCRs in dual (PDE3+PDE4)-dependent regulation of cardiac pacemaker function.

Figure 4. Dual PDE3+PDE4 activation regulates basal LCR period and characteristics.

Confocal line-scan images of representative intact SANC depicting AP-induced Ca²⁺ transients and LCRs (arrowheads) during spontaneous beating: (A) in the basal state (inset depicts measurements of the LCR size as full width at half maximum amplitude (FWHM)) and (B) during concurrent PDE3+PDE4 inhibition by 0.3 μmol/L cilostamide and 2 μmol/L rolipram. Normalized subsarcolemmal fluorescence averaged over image width is shown in black beneath the image. Insets show definitions of the LCR period, spontaneous cycle length and time to 90% decay of AP-induced Ca²⁺ transient (T-90). C, Dual PDE3+PDE4 inhibition markedly increased the LCR size (C) and (D) LCR number per each spontaneous cycle (n=7 SANC, 4 rabbits), T-test ⁺P<0.01. (E), a representative example of the decrease in the LCR periods produced by dual PDE3+PDE4 inhibition and accompanied by a decrease in the spontaneous cycle lengths of SANC in panels A and B.

Figure 5. Suppression of LCRs, but not inhibition of If current prevents an increase in spontaneous SANC firing rate by dual PDE3+PDE4 inhibition.

A, B, representative AP recordings before and after dual PDE3+PDE4 inhibition: (A) in basal conditions (rate increased from 134b/min to 185b/min); B, when RyR were inhibited by 3 μmol/L ryanodine beating rate decreased from 186 to 72b/min; dual PDE3+PDE4 inhibition increased the firing rate to only 75b/min. Right panels show superimposed spontaneous APs from left and middle panels with extended time resolution. C,D, Average acceleration of the firing rate by dual PDE3+PDE4 inhibition presented as % of baseline: C, in the absence (n=9) or presence (n=4) of ryanodine pretreatment and (D) in the absence (n=9) or presence of either 2 mmol/L Cs⁺ (n=6) or 5 μmol/L ivabradine (n=5); % of baseline was calculated relative to firing rate before treatment with the combination of cilostamide+rolipram. T-test *P<0.05.

A hyperpolarization activated “funny” current I_f regulates early part of DD and is directly activated by intracellular level of cAMP.¹⁴ To assess functional importance of I_f current for PDE-dependent regulation of spontaneous SANC firing, we compared effects of dual PDE3+PDE4 inhibition in the presence and absence of “funny” current inhibitors ivabradine²⁹ (SMFig.3) or Cs^{+ 30} (SMFig.4). Although both I_f inhibitors decreased the spontaneous SANC beating rate, acceleration of spontaneous SANC firing by dual PDE3+PDE4 inhibition was unchanged in the presence or absence of these inhibitors (SMFig.3-4, Fig.5D). This indicated that I_f current was not a target of dual PDE3+PDE4 activation.

Concerted PDE3+PDE4 activation regulated basal L-type Ca²⁺ current amplitude

L-type Ca²⁺ current generates an AP upstroke in primary pacemaker cells and provides Ca²⁺ available for pumping into SR, I_Ca,L is regulated by PDEs both in VM³¹ and SANC⁵ and could be another target of concurrent PDE3+PDE4 activation. PDE4 inhibitor rolipram did not change I_Ca,L amplitude in isolated rabbit SANC, while PDE3 inhibitor cilostamide markedly increased I_Ca,L by ∼60% (P<0.01) (Fig.6). Dual PDE3+PDE4 inhibition, however, increased I_Ca,L by ∼100% (P<0.01), an effect comparable to that of IBMX, which markedly exceeded combined effects of separate PDE3 and PDE4 inhibition (Fig.6). Thus, dual PDE3+PDE4 activation regulated basal I_Ca,L amplitude in SANC in a synergistic manner, reducing basal Ca²⁺ influx through L-type Ca²⁺ channels and limiting amount of Ca²⁺ available for pumping into SR.

Figure 6. Dual PDE3+PDE4 activation regulates basal L-type Ca²⁺ current amplitude in a synergistic manner.

A-C, representative recordings of L-type Ca²⁺ current (I_Ca,L) in response to inhibition of (A) PDE3 (0.3 μmol/L cilostamide) or (B) PDE4 (2μmol/L rolipram) alone or (C) dual PDE3+PDE4 inhibition by combination of cilostamide+rolipram. D, average increases of I_Ca,L amplitude during inhibition of PDE3 (n=7) or PDE4 (n=5) alone, concurrent PDE3+PDE4 inhibition (n=6) or IBMX (n=7)_presented as % of control. One-way ANOVA with Bonferroni's posthoc test *P<0.01 vs rolipram alone; ⁺P<0.01 vs cilostamide alone.

Concurrent PDE3+PDE4 activation regulated intrinsic SR Ca²⁺ cycling in permeabilized SANC

To study whether dual PDE3+PDE4 activation regulated intrinsic SR Ca²⁺ cycling, avoiding presence of functional ionic channels, we employed saponin-permeabilized SANC. Like in intact SANC, no changes in LCR parameters were recorded during inhibition of PDE4 alone in permeabilized SANC (SMFig.5C-D), while inhibition of PDE3 alone increased the LCR size by ∼10% (P<0.05), but changes in LCR number did not reach statistical significance (SMFig.5A-B). Dual PDE3+PDE4 inhibition, however, substantially augmented both LCR number by ∼60% (P<0.01) and LCR size by ∼25% (P<0.01) (Fig.7A-B), exceeding additive effects produced by inhibition of PDE3 or PDE4 alone. Thus, dual PDE3+PDE4 activation regulated the intrinsic SR Ca²⁺ cycling in a synergistic manner, which could be due, in part at least, to an increase in the SR Ca²⁺ load.³² To test this idea we applied a pulse of caffeine, which rapidly empties the SR Ca²⁺ store, directly on SANC. Dual PDE3+PDE4 inhibition markedly (P<0.05) increased the SR Ca²⁺ content (Fig.7C-D), confirming that augmentation of LCR parameters was partially due to an increase in the SR Ca²⁺ load.

Figure 7. Dual PDE3+PDE4 activation regulates LCR parameters and SR Ca²⁺ content in permeabilized SANC.

A, Confocal line-scan images of representative saponin-permeabilized rabbit SANC prior to and during dual PDE3+PDE4 inhibition (inset depicts measurements of the LCR size as full width at half maximum amplitude (FWHM)). B, Average changes in the LCR number (left) and size (right) produced by 2-minute superfusion with 0.3 μmol/L cilostamide and 2 μmol/L rolipram (n=6). C, Estimation of SR Ca²⁺ content by rapid application of caffeine to permeabilized SANC in the absence (top) or presence (bottom) of dual PDE3+PDE4 inhibition. D, Compared to control group (n=7) dual PDE3+PDE4 inhibition (n=14) markedly increased caffeine-induced SR Ca²⁺ release. T-test ⁺P<0.01; *P<0.05.

Combined PDE3+PDE4 activation regulated the LCR period and spontaneous SANC cycle length through modulation of the SR Ca²⁺ refilling kinetics

The increase in the SR Ca²⁺ load in response to dual PDE3+PDE4 inhibition in permeabilized SANC (without augmented Ca²⁺ supply via I_Ca,L) could be attributable, at least in part, to an increase in SERCA activity, which is regulated by cAMP-mediated PKA-dependent phosphorylation of PLB at Ser¹⁶ site.³³ Inhibition of PDE3 or PDE4 alone in rabbit SANC increased PLB phosphorylation by ∼21% (P>0.05) and ∼17% (P>0.05), respectively (Fig.8A). Concurrent PDE3+PDE4 inhibition, however, increased PLB phosphorylation by ∼108% (P<0.05), an effect that surpassed by ∼2-fold the additive effects of separate PDE3 or PDE4 inhibition and was comparable to that of IBMX (Fig.8A). Enhanced PLB phosphorylation by dual PDE3+PDE4 inhibition would increase SERCA efficiency and shorten the SR Ca²⁺ refilling time reflected in the decrease of the decay of AP-induced Ca²⁺ transient at 90% (T-90).³⁴ Indeed, there was a close link between gradations in the increase of PLB phosphorylation during separate or concurrent PDE3 and PDE4 inhibition and reduction in the SR refilling times (indexed by T-90) (Fig.8B). Reductions in T-90 during inhibition of PDE3 or PDE4 alone, concurrent PDE3+PDE4 inhibition or IBMX were replicated in reductions of LCR periods, that were closely correlated to decreases in spontaneous SANC cycle lengths (Fig.8).

Figure 8. PDE inhibition-prompted increase in PLB phosphorylation reduces SR Ca²⁺ refilling times and LCR periods, leading to decrease in spontaneous SANC cycle lengths.

A, Top, representative Western blots of total and phosphorylated PLB at Ser¹⁶ site in rabbit SANC before and during application of 0.3 μmol/L cilostamide or 2μmol/L rolipram alone, combination of cilostamide+rolipram or 100μmol/L IBMX. Bottom, average changes in P-PLB at Ser¹⁶ site expressed as % of control (n=7-9 rabbits). One-way ANOVA with Tukey's posthoc test, *P<0.05 vs cilostomide or rolipram alone. B, An increase in PLB phosphorylation was highly correlated with reduction in SR Ca²⁺ refilling times (T-90, n=4-6 SANC, 3 rabbits) and (C) proportional shortening in LCR periods, which predicted (D) decrease in spontaneous SANC cycle lengths.

Discussion

Here we established the following order of messenger RNA expression of major cAMP-degrading PDE (PDE1-4) in rabbit SANC, i.e. PDE3 and PDE4 were the dominant PDEs, with PDE3A and PDE4B as the most abundant isoforms, while expression of other PDE subtypes was significantly less (Fig.1A). Consistent with RNA data, expression of PDE3A and PDE4A protein in the rabbit left ventricle exceeded that in the SA node. Expression of PDE4B or PDE4D protein, however, was either comparable or markedly higher in the SA node, indicating that these PDEs were the most abundant PDE4 isoforms in the rabbit SA node (Fig.1B-F).

PDEs represent a unique mechanism of cAMP degradation, and targeting PDEs to specific intracellular locations might create local pools “microdomains” with high or low cAMP levels, in latter case PDEs act like “black holes” converting cAMP into 5′-AMP and thus protecting specific compartments from cAMP influx and PKA activation.³⁵ In rabbit SA node PDE3 is the dominant cAMP hydrolyzing activity representing ∼75% of total PDE activity in the SR-enriched fraction and ∼30% in the cytosolic fraction.²⁰ Compared to PDE3, PDE4 activity is less, and its major part is in cytosolic fraction ∼20% and only ∼10% in the SR-enriched fraction.²⁰ In the human ventricle both PDE3A and PDE4D associate with SR (i.e. SERCA) and play a primary role in regulation of cardiac contractility.^24,36 PDE4D is also a dominant PDE4 subtype in the human atrium, and PDE4 inhibition increases frequency of Ca²⁺ sparks and initiates Ca²⁺ waves in human atrial myocytes (AM).³⁷ Both PDE3A and PDE4D are colocalized with SERCA-PLB complex in the mouse heart.^26,36 An increase in cardiac contractility, associated with improved SERCA function and linked to augmented PLB phosphorylation, was reported in both PDE3A and PDE4D knockout mice.^26,36 Besides, PDE3A knockout mice have an increased basal heart rate.²⁶

Our study provides the first immunocytochemical evidence for the organization of major PDE3 and PDE4 subtypes in SANC. Both PDE3A and PDE4D were distributed in a striated pattern colocalized with the Z-line protein α-actinin, which did not rely on the presence of t-tubules since they are absent in SANC. Similar to human or mouse ventricular myocytes, PDE3A and PDE4D in rabbit SANC were colocalized with SR proteins SERCA and PLB (Fig. 2), suggesting that these PDE isoforms could likely regulate cAMP-mediated PKA-dependent phosphorylation of PLB in SANC. Unphosphorilated PLB inherently inhibits SERCA, but when PLB is phosphorylated by PKA it dissociates from SERCA and relieves its inhibition, promoting Ca²⁺ re-uptake and speed at which SR is refilled with Ca²⁺.³³ Compared to VM, rabbit SANC have increased amount of SERCA and reduced amount of PLB, suggesting more efficient SERCA function in the basal state.²² Our results demonstrated that dual PDE3+PDE4 activation in rabbit SANC regulated basal SERCA function in a synergistic manner through modulation of PLB phosphorylation (Fig.8A). Indeed, graded increases in PLB phosphorylation during separate or concurrent PDE3 and PDE4 inhibition were closely correlated with reductions in SR Ca²⁺ refilling times (T-90), concurrent decreases in LCR periods and spontaneous SANC cycle lengths (Fig.8). Dual PDE3+PDE4 inhibition markedly augmented LCR number and size (Fig.4) due, at least in part, to an elevation of the SR Ca²⁺content (Fig.7) and consequent synchronization of RyR Ca²⁺ release.³² The augmented RyR Ca²⁺ release beneath sarcolemma triggered elevated inward I_NCX at earlier times hastening DD rate, speeding up the occurrence of the next AP and thus increasing the spontaneous SANC beating rate (Fig.3-4, Fig.8). When RyRs were functionally disabled by ryanodine, dual PDE3+PDE4 inhibition produced only minor increase in the spontaneous SANC beating rate (Fig.5), confirming critical role of LCRs for (PDE3+PDE4)-dependent regulation of spontaneous SANC firing. Thus, concurrent PDE3+PDE4 activation operated in a synergistic manner to restrict basal spontaneous SANC beating rate through delay of LCR occurrence and suppression of LCR parameters.

L-type Ca²⁺ channels are a well-known target of cAMP-mediated PKA-dependent pathway regulated by PDE activation. Selective PDE3 inhibition markedly increased I_Ca,L amplitude both in human and rabbit AM, while PDE4 inhibition was without effect, dual PDE3+PDE4 inhibition, however, further augmented I_Ca,L producing synergistic effect comparable to that of IBMX.³⁸ Similar to human AM, dual PDE3+PDE4 inhibition in rabbit SANC worked in a synergistic manner and increased I_Ca,L amplitude by ∼100%, while inhibition of PDE4 or PDE3 alone either had no effect or moderately increased I_Ca,L by ∼60%, respectively (Fig.6). In the mouse heart PDE4B has been identified as a part of the L-type Ca²⁺ channel complex and was the major PDE isoform modulating I_Ca,L amplitude during β-AR stimulation.³⁹ In our study PDE4B and PDE3A were co-localized beneath sarcolemma of SANC (Fig.2B), suggesting that these PDE subtypes could work together restraining Ca²⁺ influx through L-type Ca²⁺ channels in a synergistic manner.

Considering that PDE inhibition increases cAMP level, we examined contribution of I_f current in the acceleration of spontaneous SANC firing by dual PDE3+PDE4 inhibition. As expected, suppression of I_f current by either ivabradine or Cs⁺ decreased spontaneous SANC beating rate, but acceleration of spontaneous firing by dual PDE3+PDE4 inhibition remained preserved regardless of I_f inhibition (Fig.5D; SMFig.3-4). Thus, regulation of spontaneous SANC firing by dual PDE3+PDE4 activation bypassed “funny” current. This result may be explained by specific location of I_f channels to lipid rafts in rabbit SANC,⁴⁰ which might provide a spatial barrier between I_f channels and increase in cAMP level produced by dual PDE3+PDE4 inhibition. This idea was supported by a recent study, which demonstrated that the positive chronotropic effect of cAMP-dependent agent glucagon was due to a cAMP-mediated activation of I_f current that was not limited by PDE3 or PDE4 activation.⁴¹

Whereas synergism between PDE3 and PDE4 has been noted before in different cell types,^8,16-19 specific mechanisms that explain synergistic effect of dual PDE3+PDE4 inhibition remains enigma. This synergistic effect could be based on co-localization of PDE3 and PDE4 (Fig.2) and specific interaction between these PDEs. Though both PDE3 and PDE4 can degrade cAMP, their affinities for cAMP are different; PDE3 degrades cAMP in the range of 10–100 nmol/L,⁴² while PDE4 degrades cAMP in the range of 2–8 μmol/L.⁴³ Therefore, in the basal state when cAMP is relatively low only PDE3 is, likely, active while PDE4 remains dormant. It was shown that PDE4 could be activated by phosphorylation,⁴⁴ e.g. PKA-dependent phosphorylation of PDE4 was associated with 2-6-fold increase in PDE4 activity.⁴⁵ Inhibition of PDE3 could be prerequisite for PDE4 activation, i.e. PDE3 inhibition increases both local cAMP and PKA-dependent phosphorylation, shifting cAMP level within the degradation range of PDE4 and concurrently activating PDE4 via PKA-dependent phosphorylation. Thus, cAMP level in the cell could be dynamically regulated by a negative feedback involving PKA-dependent stimulation of PDE4. In this case modulation of spontaneous SANC firing by PDE3 and PDE4 could be self-adaptive with full functional effect achieved only when both PDE3 and PDE4 are concurrently inhibited. Furthermore, synchronized regulation/suppression of several targets (L-type Ca²⁺ channels, PLB, etc.) by dual PDE3+PDE4 activation in a synergistic manner could be energet ically beneficial, since minor changes in local cAMP levels at multiple locations could lead to substantial functional responses.

Clinical implications of concurrent PDE3+PDE4 inhibition

Recently, dual PDE3+PDE4 inhibitors were accepted to treat patients with allergic rhinitis, asthma, and chronic obstructive pulmonary disease (COPD). In small clinical trials, dual PDE3/PDE4 inhibitor (RPL554: Verona Pharma) improved lung function without gastrointestinal side effects of “classical” PDE4 inhibitors.^46,47 Synergy of combined PDE3+PDE4 inhibition, which markedly increases drug efficacy and is beneficial in patients with asthma and COPD,¹⁵ may also produce another type of side effects. Present report revealed that spontaneous beating rate of cardiac pacemaker cells was under control of dual PDE3+PDE4 activation which operated in a synergistic manner.

Though there is no information regarding PDE subtypes in the human SA node, PDE3 and PDE4 represent major PDE activities in human AM.³⁷ The resemblance between I_Ca,L regulation by dual PDE3+PDE4 activation in human and rabbit AM³⁸ suggests that the same type of I_Ca,L regulation might exist both in rabbit (present study) and human SANC. In this case dual PDE3+PDE4 inhibition might increase spontaneous beating rate of the human heart leading to tachycardia, which could provoke atrial fibrillation. Finally, the resting heart rate in humans is associated with cardiovascular disease and death, and the risks for those increase gradually from the lowest to the highest resting heart rate values.⁴⁸ Our results suggest that changes in the human heart rate should be carefully monitored in any treatments that employ dual PDE3+PDE4 inhibitors.

Study limitations

In this report, we established that PDE3 and PDE4, working together, regulate/suppress basal spontaneous beating of cardiac pacemaker cells likely through activation of PDE3A and PDE4B/D. We also found specific targets, i.e. I_Ca,L, PLB and possibly others (beyond scope of the present study), which were modulated by dual PDE3+PDE4 activation. Additional studies, however, are needed to verify our results in human SANC as wel as elucidate not only mechanisms of PDE3 and PDE4 interactions, but also interactions of specific PDE3/4 isoforms. Unfortunately, no pharmacological tools exist currently to determine exact functional contributions of specific PDE3 or PDE4 isoforms.

What is Known?

• Basal spontaneous beating of rabbit sinoatrial node cells (SANC) is regulated by ionic currents and local subsarcolemmal Ca²⁺ releases (LCRs) from ryanodine receptors.

• SANC have elevated basal level of cAMP and cAMP-mediated PKA-dependent phosphorylation, regulated by phosphodiesterase (PDE)3 and PDE4 activation.

What the Study ADDS?

• PDE3A, PDE4B and PDE4D are the major PDE subtypes expressed in rabbit SANC; PDE3A co-localizes with a-actinin, PDE4D, SERCA and phospholamban in Z-lines, whereas PDE3A and PDE4B co-localize beneath sarcolemma.

• Basal cardiac pacemaker function is regulated by dual PDE3+PDE4 activation, which operates in a synergistic manner to limit cAMP-mediated PKA-dependent phosphorylation and to suppress LCRs, decreasing the spontaneous SANC beating rate.