ICH S7B In Vitro Assays Do Not Address Mechanisms of QT_C Prolongation for Peptides and Proteins - Data in Support of Not Needing Dedicated QT_C Studies

Current cardiac safety testing focuses on detecting drug-induced QT(C) prolongation as a surrogate for risk of Torsade de Pointes. The nonclinical strategy, described in International Conference on Harmonization (ICH) S7B, includes in vitro assessment of hERG block or ventricular repolarization delay and in vivo QT prolongation. Several studies have reported predictive values of ICH S7B results for clinical QT(C) outcomes for small molecules; none has examined peptides and proteins other than monoclonal antibodies. To address this knowledge gap, information for peptides and proteins submitted to the US Food and Drug Administration (FDA) was collected. Results of hERG assays, ventricular repolarization assays, and in vivo QT assessment were compared with clinical QT(C) study outcomes. The results show that 14% clinical QT(C) studies for approved and investigational products failed to exclude 10-ms QT(C) prolongation. Clinical QT(C) prolongation for these molecules lacked concentration-dependence which is expected for hERG block-mediated mechanism or QT(C) prolongation could not be excluded due to characterization in the clinical study. The hERG and ventricular repolarization assays do not identify clinical QT(C) prolongation potential for peptides and proteins. Lack of alignment between hERG and ventricular repolarization assay results and clinical QT(C) outcomes suggests that the mechanisms of QT(C) prolongation by some peptides and proteins are unrelated to direct cardiac ion channel block. Similar to large targeted proteins and monoclonal antibodies, peptides and proteins regardless of size have a low likelihood of direct cardiac ion channel interactions. This characteristic supports waiving the requirement for thorough QT assessment for products comprised of naturally occurring amino acids unless proarrhythmia potential is suggested by nonclinical or clinical data.