Atrial fibrillation (AF) is the most common sustained arrhythmia, yet the mechanisms that lead to AF are incompletely understood. The occurrence of atrial fibrillation is often associated with haemodynamic or mechanical disorders of the heart, which lead to intra-atrial pressure or volume overload. Therefore myocardial stretch might play an important role in the development of AF. This is also emphasized by the common clinical finding that atrial dilatation correlates with the susceptibility to atrial fibrillation. The electrophysiological changes that occur in response to mechanical perturbations are referred to as mechanoelectrical feedback, a phenomenon that has been recognized in both experimental settings and humans. Atrial stretch caused by ventricular contraction modulates the atrial flutter cycle length in man, while in isolated hearts, acute atrial dilatation facilitates the induction and maintenance of atrial fibrillation.

We hypothesized that stretch-activated channels (SACs) might be involved in the arrhythmogenic effect of acute atrial stretch. SACs are found in many cardiac tissues and species including humans. Non-selective cationic SACs pass Ca²⁺ as well as Na⁺ and K⁺, whereas others selectively carry K⁺ and possibly Cl^–. Block of SACs has previously shown to decrease arrhythmic responses to acute stretch in ventricular myocardium.

We studied the effect of SAC on the inducibility of atrial fibrillation in the isolated Langendorff perfused rabbit heart. In the unstretched atrium at a pressure of 0 cmH₂O, burst pacing did not produce AF. With an increase in atrial pressure, AF could be induced in all preparations. At low pressures, the sinus rhythm recovered spontaneously. At higher pressures, AF was sustained. AF inducibility increased and average duration of AF episodes lengthened as a function of pressure. Sustained AF terminated promptly upon lowering the intra-atrial pressure. We demonstrated that Gadolinium (Gd³⁺), a generic blocker of stretch-activated channels, reduced the ability of stretch to potentiate atrial fibrillation in a dose-dependent manner. After application of Gd³⁺ (50 mM) in 16 hearts, intra-atrial pressure needed to be raised to 19.0 ± 0.5 cmH₂O to allow perpetuation of AF (vs. 8.8 ± 0.2 cmH₂O at baseline, P < 0.001).

SACs have been inhibited by a variety of non-specific agents like Gd³⁺, streptomycin and amiloride in patch-clamp experiments. Their lack of specificity prevents clinical applications as blockers of SACs. It was recently reported that spider venom from the tarantula Grammostola spatulata could block SACs in single myocardial cells and very recently the active agent from the venom, GsMtx-4, was purified. It is a 4 kD peptide of the cysteine knot family that specifically blocks SACs as well as volume-activated cation currents in rat primary astrocytes and in hypertrophic dog ventricular cells.

After application of 170 nM GsMtx-4 in the isolated rabbit heart (n = 10), induction of AF required a significantly higher atrial pressure: 18.5 ± 0.5 vs. 8.8 ± 0.4 cmH₂O at baseline (P < 0.001). Sustained AF was obtained at 24.8 ± 0.6 cmH₂O after GsMtx-4 vs. 11.6 ± 0.3 cmH₂O at baseline (P < 0.001). The average duration of AF decreased at intra-atrial pressures between 10 and 27.5 cmH₂O (P < 0.05). These GsMtx-4-related effects were readily reversible upon wash-out.

To study the effect of SAC blockade on atrial excitability, the right atrial effective refractory period (ERP) was determined (n = 7). ERP progressively shortened with an increase in atrial pressure. On average, the ERP shifted from 71 ± 7 ms at 0.5 cmH₂O to 39 ± 2 ms at 20 cmH₂O (P < 0.05). GsMtx-4 did not significantly alter the ERP at any pressure level. This was in accordance with previous findings during Gd³⁺ application.

Our results demonstrate that Gd³⁺ and GsMtx-4, the only known specific blocker of SACs, reduce the vulnerability to atrial fibrillation during acute atrial dilatation. They confirm the predicted effects of SAC blockade on stretch-induced arrhythmias. In our model, electrical stimulation could only elicit atrial fibrillation when the atrium was preconditioned by stretch, while increased dilatation also increased the duration of AF. SAC blockade decreased the probability of initiation and increased the probability of spontaneous recovery of AF. This result suggests the involvement of SACs in the process.

The present data reaffirm that myocardial stretch shortens refractoriness in the atria. According to the multiple wavelet hypothesis, AF represents multiple wavefronts encountering excitable tissue. Shortening of refractoriness shortens the excitation wavelength and thereby favours re-entrant activity. Yet, after SAC blockade we found that local atrial refractoriness was not significantly altered while vulnerability to AF was markedly reduced. This implies that GsMtx-4 is acting at specific sites that initiate AF rather than by simply lengthening the ERP through generic changes in excitability. Importantly for potential clinical applications, GsMtx-4 had no significant side-effects on the preparation, thereby opening the field for a new class of anti-arrhythmic agents. The availability of a specific reagent to modify SACs will permit study of the coupling between mechanical strain and electrical excitability in further detail.