Heart failure-induced atrial remodelling promotes electrical and conduction alternans

Fig 1. Effect of HF-induced electrical remodelling on the APD90, CaT amplitude and SR Ca2+ content. (Ai), (Bi) and (Ci) Recordings of APs, intracellular Ca2+ concentration ([Ca2+]i) and Ca2+ concentration in NSR ([Ca2+]NSR) under control and HF conditions at 1Hz. (Aii) and (Bii) The increased APD90 (ΔAPD90) and increased CaT amplitude (ΔCaT amplitude) under HF conditions relative to control from experimental data [18] (grey and black) and simulations (while) at 1 Hz and 2Hz. (Cii) The increased SR Ca2+ content (ΔSR Ca2+ content) in HF relative to the control from experimental data [18] (grey and black) and simulations (while) at 1 Hz.

Fig 2. Relative contributions of 7 parameters to the APD90, CaT amplitude and SR Ca2+ content using parameter sensitivity analysis at 1 Hz. (A) APD90. (B) CaT amplitude. (C) SR Ca2+ content.

Fig 3. Effect of HF-induced electrical remodelling on rate dependent curves and the restitution curves. (A) APD90 rate dependent curves, (B) CaT amplitude rate dependent curves, and (C) CaT decay time rate dependent curves from two consecutive beats in a quasi-steady state under control (shown in blue) and HF (shown in orange) conditions. (D) APD90 restitution curves at a quasi-steady state under control (blue) and HF (orange) conditions. (E) Slopes and (F) maximum slope of the restitution curves in (D). (ii) Enlarged view of (i) to more clearly reveal the PCL range with alternans.

Fig 4. Traces of AP, Ca2+ transient, JSR Ca2+ concentration and ion channel currents during long-short alternans at a PCL of 148 ms. (A) AP. (B) [Ca2+]i. (C) JSR Ca2+ concentration ([Ca2+]JSR). (D) INa. (E) ICa. (F) Jrel. (G) Ito. (H) INCX. (I) Jup.

Fig 5. Sensitivity analysis of alternans to HF-induced remodelling in HF model. (A) Each parameter remodelling was removed from the HF model one at a time compared with control and HF models. (B) Kup was reduced from 0% to 90% relative to the control in the HF model. (C) Gto was reduced from 0% to 90% relative to the control in the HF model. (D) GCa was reduced from 0% to 60% relative to the control in the HF model.

Fig 6. 1D simulation results under control and HF conditions. (A) Normal conduction, discordant alternans, and 2:1 conduction block in the D-PCL parameter space under control (i) and HF (ii) conditions. (B) The rate dependent curves of CV under control (i) and HF (ii) conditions, as measured by implementing three-member cell-to-cell coupling (D = 0.167 mm2/ms, 0.084 mm2/ms, and 0.028 mm2/ms). (Bii) is an enlarged view of (Bi) that more clearly reveals the 130 ms to 200 ms PCL range.

Fig 7. Representative 1D simulation of the discordant alternans with D = 0.028 mm2/ms at a PCL of 168 ms in HF. (A(i)) Time-space plot of the membrane potential of a homogeneous, 1D cable. (A(ii)) The corresponding APD90 of each beat, in which black dashed lines with arrows indicate the APD node position as the beat number increases. (A(iii)) Time course traces of AP for the 40th, 100th, and 170th cells, which are marked with black lines in (A(i)). (B(i)) Time-space plot of the Ca2+ transient in a homogeneous, 1D cable. (B(ii)) The corresponding peak Ca2+ transient of each beat, in which black dashed lines with arrows indicate the peak Ca2+ transient node position as the beat number increases. (B(iii)) Time course traces of Ca2+ transient for the 40th, 100th, and 170th cells, which are marked with black lines in (B(i)).

Fig 8. Effect of cell-to-cell coupling on the node position and number with PCL = 168 ms under the HF condition. (A) Results with D = 0.167 mm2/ms. (B) Results with D = 0.028 mm2/ms. (i) APD90 and (ii) DI spatial distribution for consecutive beats.