Impact case study database

All new pharmaceutical compounds must be carefully screened for unwanted side effects on the heart which can cause fatal arrhythmias. Many drugs have been withdrawn from the market worldwide because of fatalities due to these arrhythmias at a cost of billions of dollars. Even drugs for ailments as mild as hay fever have been withdrawn due to fatalities and an unacceptable increase in arrhythmic risk (e.g. terfenadine), so there is now a huge industrial effort in avoiding this situation occurring for new drug compounds. The root cause has been identified as drug molecules binding to and blocking ion channels in the heart.

Ion channel block can now be screened in immortal cell lines in a dish ( in vitro testing) at an early stage of drug development without using cardiac muscle tissue from animals. This is the standard screening approach for a number of ion channels in many large pharmaceutical companies. However, clinical risk is a complex function of the degree of block of many cardiac ion channels which is difficult to understand by looking at results of these screens independently.

Innovatively, the mathematical modelling approach developed by Mirams offers a way to integrate information from different ion channel in vitro screens, to predict the effect on whole cardiac muscle cells. The resulting simulations can provide a more accurate summary of the risk of a new potential drug compound earlier in drug development and hence more cheaply than the pre-existing approaches. In short, the computational modelling tool, ApPredict, uses the data produced by in vitro tests to identify lower-risk compounds to be progressed into translational in vivo studies on animals (with a concomitant reduction in the need for animal studies) and ultimately into clinical trials on humans.

Industrial Use of Electrophysiology Simulations

Impact on ApPortal users:

Mirams has been developing the Chaste-based cardiac simulation software known as ‘ApPredict’ [1, 2, 6-9] in partnership with a web developer to embed it in a web interface (ApPortal) since 2011. The ApPredict software takes information on the extent to which drugs block particular ion channels, then provides a prediction for the integrated effect on the whole cardiac muscle cell.

Improvements by Mirams to ApPredict whilst at Nottingham (2016-onwards) led to the 2019.1 version of ApPortal being made available via a public website where anyone can register for an account and run cardiac safety simulations after manually entering data on the degree to which compounds block specific ion channels. The portal is very widely used and has performed a total of over 7000 simulations for over 350 users. Based on email domain usernames, there are users from 147 companies including Contract Research Organisations, small pharmaceutical companies and 14 of the top 20 largest global pharmaceutical companies (by 2019 revenue); and over 70 educational or research establishments [A].

Impact on the SME Bioprojet-BIOTECH:

In 2017, Mirams worked with Bioprojet-BIOTECH to assess the pro-arrhythmic risk of their new compound ‘pitolisant’ [3]. The drug is now marketed as ‘Wakix’ for treatment of narcolepsy, a condition estimated to affect 1 in 2500 people. Bioprojet states that Mirams’ simulation tool led to enhanced confidence in their understanding of the compound action and hence in their assessment of its pro-arrhythmic risk potential [D]. They also acknowledge the contribution of the research [3] in obtaining FDA approval [E] for ‘pitolisant’ [D].

Impact on GlaxoSmithKline and Roche:

Since Mirams’ return to Nottingham in 2016, he has carried out development on the underlying simulation software, funded on a consultancy basis by GlaxoSmithKline [B] and Hoffman-LaRoche [C]. During the same period, the ApPortal has been extended to provide uncertainty quantification features (uncertainty in model predictions due to the variability associated with the input data), and these are now in use within the companies and public portal [A, B, C]. The portal has been deployed inside company firewalls at GSK and Roche where it accesses their safety screening databases directly to acquire input data for the simulations, with the user simply typing in an internal compound ID rather than individual screening results. This system is in regular use by Safety Pharmacology teams at both companies in routine safety testing and compound profiling [B, C].

The Vice President, Non-Clinical Safety at GlaxoSmithKline reports:

“ApPredict v2019.1 developed at the University of Nottingham (based on Chaste v2019.1 discussed in Cooper et al. [5] *) is now in use at GSK to predict the effects of multiple ion channel block on various cell types: stem-cell derived myocytes, rabbit ventricular myocytes and adult human myocytes.*” [B]

Similarly, a Principal Scientist of Pharma Research and Early Development at Roche reports:

“The simulator is installed within Roche to predict the effects of multiple ion channel block on stem-cell derived myocytes, guinea pig myocytes and adult human myocytes. We are using the July 2020 version of ApPredict, based on Chaste v2019.1 (Cooper et al. [5] *).*” [C]

Both pharmaceutical companies use the portal to (i) screen their compounds and prioritise which to take forward (to advance compounds that are less likely to fail later safety tests), leading to enhanced productivity and cost-saving benefits; (ii) to assist in design of later safety tests (to select drug concentrations of most interest and relevance); and (iii) to compare predictions with later safety test results to check whether all findings are consistent and that the drug’s action is well understood [B, C]. Finally, the worldwide population will be less likely to be given drugs that can cause arrhythmia, and there are significant animal reduction, refinement, replacement (3Rs) benefits [B, C].

For example, GSK reports benefits to their business:

“*Having implemented ApPredict […] allowed GSK to enhance our internal cardiovascular safety strategy and to comply with the shifting landscape in pharmaceutical drug development.*” [B]

and also the 3Rs benefits to the approach:

“Since its implementation in 2017 as part of the established early cardiovascular screening strategy, all small molecules […] progressing to clinical candidates have been evaluated by the ApPredict model, leading to a >90% reduction in animal use” [B].

Meanwhile, Roche reports:

“[ApPredict] has been successfully applied in multiple (>10) projects across different therapeutic areas and developmental stages to optimize safety profile of clinical candidates. The main impact included the reduction in the number of animal studies with tens of animals per compound, optimization of the clinical study design by lowering the amount of ECG monitoring and potential waiving of the phase III thorough QT study” [C].

Provision of Training in Simulation Approaches to Pharma To encourage knowledge exchange between mathematicians and the pharmaceutical industry Mirams has run workshops for non-mathematician safety pharmacologists to learn how to run cardiac electrophysiology simulations, for example as part of a Continuing Education course for the Safety Pharmacology Society for approximately 30 industrial safety pharmacologists [F]. Mirams was an Academic Mentor for Dr Ken Wang based in Roche, Basel, Switzerland on a Roche Postdoctoral Fellowship 2015-2017 (she now has a permanent post there), and has also jointly supervised PhD students (2013-present) with Roche [C]. Mirams has a blog featuring accessible articles on Safety Pharmacology topics ( Mathematical Matters of the Heart, over 15,000 page views in 2017-2020 as accessed 13 January 2020); the majority of its readers are from the USA (where many large pharma companies are based), followed by UK, Germany, India, France, South Korea, Australia, Canada and the Netherlands [G].