Late stage failure of new therapies during clinical trials is a huge concern to the pharmaceutical industry due to the significant costs involved in reaching this stage in the development process, and also the potential safety risks to patients.
BMS-986094 was developed as a prodrug of a guanosine nucleotide analogue developed to treat the Hepatitis C virus. It was discontinued in Phase 2 clinical trials due to cardiotoxicity – with one death and eight patients hospitalised due to significantly reduced left ventricular ejection fraction (LVEF). Further analysis of ECGs of the patients with LVEF indicated ST depressions, T-wave inversions or loss of T-wave amplitude. Interestingly, the cardiotoxic effect was only observed after extended use following administration of the 200 mg dose. Previous studies have indicated that inhibition of mitochondrial RNA polymerase may be responsible for the adverse effects.
Since the discontinuation of the development of BMS-986094 in 2012, in vitro methodologies for detecting such liability have improved significantly. Techniques such as iPSC-derived cardiomyocytes and microelectrode array are now being routinely used for preclinical screening, and form an integral part of the CiPA initiative to overhaul preclinical cardiac safety testing.
Our research at Cyprotex has focused on investigating further the mechanism(s) behind BMS‑986094 cardiotoxicity using some of these more sophisticated techniques. Human iPSC-derived cardiomyocytes were selected for these studies due to their ability to maintain viability in long term culture so chronic dosing could be replicated. Over the course of 14 days, a number of endpoints were analysed following exposure to BMS-986094. These included MEA, calcium flux and mitochondrial biogenesis. By using these techniques, our understanding of the toxic effects of BMS-986094 has been enhanced significantly.
In terms of the MEA, no effect was observed at 1 hr indicating that the adverse effects only occurred after chronic long term dosing. By 14 days, loss of beating was observed at 0.4 µM, 2 µM and 10 µM. Even at 80 nM, beat rate and sodium amplitude increased significantly at the later time points, demonstrating the sensitivity of the MEA recording. Calcium flux was monitored on a Hamamatsu FDSS/µCell, and corroborated the earlier MEA data, with loss of calcium flux at the top two concentrations and very low level calcium flux at the 0.4 μM level. The beat rate for the 80 nM concentration was recapitulated in this assay. Finally, high content screening (HCS) was used to assess mitochondrial biogenesis. The data suggest that BMS-986094 does not block expression of the mitochondrial coded proteins and actually may cause the cell to respond by upregulating mitochondrial protein. Based on the timing and expression of the mitochondrial protein, it is likely that the mitochondrial effect is unrelated to the toxicity observed.
Our research demonstrates the value of using multiple in vitro assay platforms over an extended time course to mimic chronic dosing regimens. Using this approach, the detection of cardiac liability is enhanced greatly. Designing similar screening strategies for new therapies currently in preclinical discovery and development would result in a significantly improved safety profile for compounds progressing into animal testing and human clinical trials.
This research was presented at SOT in San Antonio Texas from 11th-15th March 2018.