TBX18 overexpression enhances pacemaker function in a rat subsidiary atrial pacemaker model of sick sinus syndrome

Abstract

Key points: The sinoatrial node (SAN) is the primary pacemaker of the heart. SAN dysfunction, or 'sick sinus syndrome', can cause excessively slow heart rates and pauses, leading to exercise limitation and syncope, currently treated by implantation of an electronic pacemaker. 'Biopacemaking' utilises gene therapy to restore pacemaker activity by manipulating gene expression. Overexpressing the HCN pacemaker ion channel has been widely used with limited success. We utilised bradycardic rat subsidiary atrial pacemaker tissue to evaluate alternative gene targets: the Na⁺ /Ca²⁺ exchanger NCX1, and the transcription factors TBX3 and TBX18 known to be involved in SAN embryonic development. TBX18 overexpression restored normal SAN function, as assessed by increased rate, improved heart rate stability and restoration of isoprenaline response. TBX3 and NCX1 were not effective in accelerating the rate of subsidiary atrial pacemaker tissue. Gene therapy targeting TBX18 could therefore have the potential to restore pacemaker function in human sick sinus syndrome obviating electronic pacemakers.

Abstract: The sinoatrial node (SAN) is the primary pacemaker of the heart. Disease of the SAN, sick sinus syndrome, causes heart rate instability in the form of bradycardia and pauses, leading to exercise limitation and syncope. Biopacemaking aims to restore pacemaker activity by manipulating gene expression, and approaches utilising HCN channel overexpression have been widely used. We evaluated alternative gene targets for biopacemaking to restore normal SAN pacemaker physiology within bradycardic subsidiary atrial pacemaker (SAP) tissue, using the Na⁺ /Ca²⁺ exchanger NCX1, and the transcription factors TBX3 and TBX18. TBX18 expression in SAP tissue restored normal SAN function, as assessed by increased rate (SAN 267.5 ± 13.6 bpm, SAP 144.1 ± 8.6 bpm, SAP-TBX18 214.4 ± 14.4 bpm; P < 0.001), improved heart rate stability (standard deviation of RR intervals fell from 39.3 ± 7.2 ms to 6.9 ± 0.8 ms, P < 0.01; root mean square of successive differences of RR intervals fell from 41.7 ± 8.2 ms to 6.1 ± 1.2 ms, P < 0.01; standard deviation of points perpendicular to the line of identity of Poincaré plots (SD1) fell from 29.5 ± 5.8 ms to 7.9 ± 2.0 ms, P < 0.05) and restoration of isoprenaline response (increases in rates of SAN 65.5 ± 1.3%, SAP 28.4 ± 3.4% and SAP-TBX18 103.3 ± 10.2%; P < 0.001). These changes were driven by a TBX18-induced switch in the dominant HCN isoform in SAP tissue, with a significant upregulation of HCN2 (from 1.01 × 10^-5 ± 2.2 × 10^-6 to 2.8 × 10^-5 ± 4.3 × 10^-6 arbitrary units, P < 0.001). Biophysically detailed computer modelling incorporating isoform-specific HCN channel electrophysiology confirmed that the measured changes in HCN abundance could account for the observed changes in beating rates. TBX3 and NCX1 were not effective in accelerating the rate of SAP tissue.

Keywords: TBX18; biopacemaking; gene therapy; sick sinus syndrome; subsidiary atrial pacemaker tissue.

Figure 1. Adenovirus‐mediated expression of TBX18, but not NCX or TBX3, increases the beating rate of the SAP preparation

A, uninfected SAP beating rates were significantly slower than the SAN. B, in SAP‐TBX18 beating rates diverged from uninfected SAP tissue after 20 h of culture, and were significantly faster than SAP at the final analysis (F). The rates during the control period were not significantly different (E). C and D, there was no significant effect on the rate when the SAP preparation was injected with Ad‐TBX3 or Ad‐NCX1‐GFP. Statistical significance displayed in panels A–D refers to the pre‐specified comparison of the final 2 h rates between treated and untreated SAP by multiple ANOVA. Statistical significance in E and F refers to comparison made to the rate of the untreated SAP preparation. G and H, expression of TBX18 and TBX3 was validated by qPCR, showing significant upregulation in treated versus untreated SAP preparations. Comparisons were also made to SAN preparations. SAN, n = 15; SAP, n = 14; SAP‐TBX18, n = 8; SAP‐TBX3, n = 9; SAP‐NCX1, n = 6. NS, not significant; ^** P < 0.01; ^*** P < 0.001. Dotted lines in B–D represent SEM. Individual data points are shown by dots in bar graphs.

Figure 2. TBX18 improves heart rate stability in the SAP

A–C, compared to untreated SAP there was a significant improvement in all measures of heart rate stability studied in SAP‐TBX18. D, Poincaré plots displayed fewer clusters of outlying short‐coupled RR intervals or pauses in SAP‐TBX18 preparations, though this did not reach significance. These premature beats (such as the example shown in Eb) demonstrated a change in morphology suggesting an ectopic focus of origin. E, examples of heart rate behaviour in the SAN; Ea, example extracellular potential recording from SAN preparation that showed rate instability with sudden changes in RR intervals; Eb, example Poincaré plot from a SAN preparation that displayed heart rate instability with high SD1 and outlying clusters representing short RR intervals and pauses. F, examples of unstable heart rate behaviour in SAP tissue; Fa, example RR plot from a SAP preparation that showed rate instability with sudden heart rate decelerations; Fb, example Poincaré plot from a SAP preparation that displayed heart rate instability with high SD1 but without outlying clusters. G, examples of stable heart rate behaviour in SAP‐TBX18 preparations; Ga, example extracellular potential recording from a SAP‐TBX18 preparation that had a stable rate and minimal variation in RR intervals; Gb, example Poincaré plot from SAP‐TBX18 preparation that displayed a stable heart rate with low SD1 and no outlying clusters. SAN, n = 10; SAP, n = 14; SAP‐TBX18, n = 8. NS, not significant; ^* P < 0.05; ^** P < 0.01. Individual data points are shown by dots in bar graphs.

Figure 3. Comparison of mRNA levels of HCN isoforms

A–C, relative abundance of HCN transcripts as determined by qPCR in the SAN, SAP and SAP‐TBX18. D–F, ratios of HCN isoforms in SAN, SAP and SAP‐TBX18. Levels are shown relative to HCN4, which is the most abundant isoform in the SAN. TBX18 causes a change in the relative levels of the isoforms in the SAP preparation with a significant upregulation of HCN2 (F) compared to the untreated SAP (E). SAN, n = 8; SAP, n = 8; SAP‐TBX18, n = 8. NS, not significant; ^* P < 0.05; ^** P < 0.01; ^*** P < 0.001. Individual data points are shown by dots in bar graphs.

Figure 4. Relative abundance of further ion channels relevant to the normal sinus node as determined by qPCR in SAP and SAP‐TBX18

Only RYR2 demonstrated a significant rise in SAP‐TBX18 compared to untreated SAP. There were no significant changes in the other measured genes. SAN, n = 8; SAP, n = 8; SAP‐TBX18, n = 8. NS, not significant; ^* P < 0.05. Individual data points are shown by dots in bar graphs.

Figure 5. Biophysically detailed computer modelling based on measured HCN mRNA abundances

A and B, computed membrane potentials (Aa and Ba) and I _f (Ab and Bb) are shown. The SAN (black lines) was compared to SAP (red lines, Aa and Ab) and to SAP‐TBX18 (red lines Ba and Bb). C, simulated heart rates; SAP‐TBX18 (HCN only) refers to the simulation only accounting for changes to HCN channel levels; SAP‐TBX18 (HCN+RYR2) refers to the simulation accounting for changes to HCN channel levels and RYR2 levels. D, simulated current–voltage relationships for each preparation normalised to cell capacitance. [Color figure can be viewed at wileyonlinelibrary.com]