In silico modelling
A population of n=500 human cardiac action potentials (AP) were simulated using the endocardial configuration of the ToR-Ord model (17) run in the Brain Dynamics Toolbox (18).To incorporate population variability in expression levels of the ion channels and electrogenic transporters that regulate cardiac excitability (i.e. the rhythmonome genes), the maximum conductance for each current (L-type calcium current, sodium current, transient outward current, late sodium current, rapid delayed rectifier (hERG) potassium current, slow delayed rectifier potassium current, inward rectifier potassium current, background potassium current, sodium calcium exchanger, sodium potassium pump, background sodium current, background calcium current, calcium pump, chloride current through calcium channel, background chloride current, calcium release from SR, calcium uptake to SR; Supplemental table 1) was multiplied by a conductance scalar (G x) randomly drawn from a log-normal distribution (19, 20) with unit mean and variance of 0.05, which we previously demonstrated can account for the range of QT intervals measured from normal and Long QT patient populations (21). Hypokalemia was simulated by manipulating parameterko in the model. The pharmacology of hydroxychloroquine were represented by an additional set of scaling factors, informed by the IC50 measured in our hERG screen, as well as data describing block of inward rectifier potassium current (IK1), cardiac sodium current (INa), and the cardiac L-type calcium current (ICaL) from the literature (22) (Supplemental table 2). Simulated myocytes were paced at 1Hz and allowed to equilibrate for 300 beats prior to analysis. Source code: https://git.victorchang.edu.au/projects/CC/repos/.