Figures
Figure 1. Effects of HCQ on ion currents attributable to hNav1.5
(INa, A), Cav1.2 (ICa,L, B), Kv4.3
(Ito, C ), hERG (IKr, D), KCNQ1/E1
(IKs, E) and Kir2.1 (IK1, F). (A-F) Left
panels (i): Typical currents in the presence of 0, 1, 10 and 100 µM HCQ.
Insets summarize the voltage clamping protocols. Right panels (ii): Mean
concentration-response plots constructed for the maximum magnitude of
each current (n cells).
Figure 2. Effects of AZM on ion currents attributable to hNav1.5
(INa, A), Cav1.2 (ICa,L, B), Kv4.3
(Ito, C ), hERG (IKr, D ),
KCNQ1/E1 (IKs, E) and Kir2.1 (IK1, F).
(A-F) Left panels (i): Typical currents in the presence of 0, 1, 10 and
100 µM HCQ. Insets summarize the voltage clamping protocols. Right
panels (ii): Mean concentration-response plots constructed for the
maximum magnitude of each current ((n cells).
Figure 3. Effects of HCQ in combination with 10 µM AZM on ionic currents
attributable to hNav1.5 (INa, A), Cav1.2
(ICa,L, B), Kv4.3 (Ito, C ), hERG
(IKr, D), KCNQ1/E1 (IKs, E) and Kir2.1
(IK1, F). (A-F) Left panels (i): Typical currents in the
presence of 0, 1, 10 and 100 µM HCQ with 10 µM AZM. Insets summarize the
voltage clamping protocols. Right panels (ii): Mean
concentration-response plots constructed for the maximum magnitude of
each current (n cells).
Figure 4. Computational docking studies. (A) hERG homotetramer structure
(PDB ID: 5va2) with HCQ molecule (space-filling representation)
indicating binding site placement. Residues 544-671 used in IFD of hERG
chains A, B, C, and D colored red, green, yellow, and blue,
respectively. (B) Overlay of IFD HCQ positions. (C) Representative,
top-scoring positions of HCQ (ball-and-stick). Y652 and F656 residues in
stick representation colored accordingly to their parent chains. (D)
2-dimensional (2D) diagram depicting interaction types between presented
HCQ position and hERG. (E) Representative, top-scoring position of AZM,
and (F) 2D diagram depicting AZM-hERG interactions. Diagrams were
prepared using Non-bond Interactions Monitor tool in Discovery Studio
Visualizer (Dassault Systèmes, Vélizy-Villacoublay, France). IFD =
Induced Fit Docking.
Figure 5. Multi-electrode array (MEA) mapping and electrocardiographic
(ECG) studies in isolated perfused guinea-pig hearts following HCQ and
AZM challenge. (A) Schematic summary of stimulation and MEA and ECG
recording configuration. (B) Typical left atrial and left ventricular
isochronal conduction maps obtained in the presence of 0, 0.01, 0.1, 1.0
and 10 µM HCQ. (C) (i) Successive LV isochronal maps before and
following addition of 10 µM HCQ and 10 µM HCQ and 10 µM AZM in
combination followed by drug washout. (ii) Quantified conduction
velocities through these conditions. (D) ECG studies: (i) ECG records
obtained before and following application of 10 µM HCQ, and 10 µM HCQ
and 10 µM AZM combined, followed by drug washout. (ii)-(v):
corresponding measured (ii) heart rates (HR), (iii) QT and (iv) PR
intervals and (v) QRS durations (n hearts) * p<0.05, **
p<0.01, *** p<0.001
Figure 6. Voltage RH237 and Rhod-2 optical mapping in isolated perfused
hearts. Comparisons of results before and following challenge by 1 µM
and 10 µM HCQ, and 10 µM HCQ combined with either 1 or 10 µM AZM. (A, B)
RH237 mapping: (A)(i) Optical mapping and ECG monitoring configuration
(ii) Maps of action potential initiation and conduction; (iii) heart
rates and (iv) conduction velocities. (B) Action potential time-courses
and recoveries: (i) Maps of AP durations at 90% repolarization
(APD90). (ii) Representative APs recorded from defined
regions of interest. (iii) APD90s averaged over the
field of view. (C) Rhod-2 AM (CaD) mapping of ventricular
Ca2+ transients (CaT): (i) maps of CaD durations at
90% recovery (CaD90). (ii) Comparison of typical CaT
transients averaged over defined regions of interest; (iii)
CaD90 magnitudes; (iv) CaD90 dispersions
(n hearts). NS not significant, * p<0.05, ** p<0.01,
*** p<0.001.
Figure 7. Voltage RH237 optical mapping of re-entry and ventricular
arrhythmic tendency in isolated perfused hearts. (A) Recording and
programmed stimulation configuration. (B) Optical action potential
traces obtained under programmed pacing at 190 ms, 180 ms and 170 ms
cycle lengths (CLs), before (control) and following addition of 10 µM
HCM, and 10 µM HCM with added 10 µM AZM. (C) Corresponding voltage maps
obtained at the numbered colour-coded timepoints (n hearts).
Figure 8. In silico modelling of action potentials before and following
HCQ and combined HCQ and AZM challenge. (A) Simulated action potentials
of human ventricular cell under control conditions and in the presence
of HCQ (10 µM) and HCQ (10 µM) + AZM (1 µM) combined. HCQ and AZM when
combined produced a synergistic action in prolonging action potential
duration (APD90), as well as reducing the repolarization
rate following the end of phase 0 depolarisation. (B) Effects of HCQ and
HCQ + AZM combined on APD90 and the maximal upstroke
velocity (MUV) of the AP. (C) Illustration of electrophysiological
effects and possible mechanisms underlying the pro-arrhythmic effects of
applications of HCQ and HCQ and AZM combined at the ion channel,
cellular and tissue levels.