“We have established an approach that transforms real-world, in-human treatment outcomes and safety data into extensive datasets to better characterize desirable and undesirable interactions between molecular pathways and drugs,” explains Dr. David Jackson, Chief Innovation Officer at Molecular Health and co-author on all three articles. “The studies demonstrate a standardized strategy to extend mechanistic modeling and systems pharmacology into drug safety and mode of action assessments that has relevance for drug development and a variety of other contexts.”

In an article published in CPT: Pharmacometrics & Systems Pharmacology, the research groups at Molecular Health and University of Florida discuss the principles and future potential of applying a reverse translational strategy in drug development to investigate drug effects at the molecular level, understand causality, and anticipate prevalence of adverse effects. They then demonstrate the approach in two subsequent articles authored with collaborators at Takeda Oncology and thinkQ². By mapping population-level adverse drug reaction data extracted from the U.S. Food and Drug Administration (FDA) Adverse Event Reporting System back to molecular pathways and target profile data, the two proof-of-concept studies examined potential causes of cardiotoxicity and of colitis associated with well-known targeted cancer therapies. Using this analytical method, the authors developed hypotheses about different molecular mechanisms that accentuate or attenuate adverse effects depending on the nature of drug-drug combinations or on a drug’s target.

Dr. Sara Kim, Assistant Professor in the Department of Pharmaceutics, and Professor Emeritus Dr. Lawrence Lesko led the research collaboration. With nearly 20 years of experience directing Clinical Pharmacology at the U.S. FDA, Dr. Lesko is a renowned systems pharmacologist and Founding Director of the Center for Pharmacometrics and Systems Pharmacology at the University of Florida in Lake Nona. He emphasizes the importance of this work: “Registries of adverse events constitute a diverse, massive and valuable data source that has not been tapped into to generate and assess molecular hypotheses about causality. The expansion of adverse event information coupled with biological and chemical data, as demonstrated in these studies, is an improvement over traditional pharmacovigilance and can inform future clinical trial designs and personalized medicine.”

Merging molecular knowledge with post-marketing adverse event reports to assess drug safety

The first proof-of-concept study, published in Clinical and Translational Science (CTS), analyzed mechanistic hypotheses in light of cardiotoxicity reports from patients on trastuzumab (Herceptin^R) combined with one of four other medications. A monoclonal antibody, trastuzumab targets HER2 and is used to treat breast cancer and advanced stomach cancer. Corroborated with experimental findings in the literature, the study suggested that trastuzumab-induced cardiotoxicity may be enhanced by doxorubicin via mitochondria dysfunction in cardiomyocytes, whereas the combination with tamoxifen, paroxetine, or lapatinib may diminish the undesirable effect by increasing antioxidant activities.

The second proof-of-concept study, published in CTS: Clinical and Translational Science, investigated autoimmunity associated with immunotherapies. The immune checkpoint inhibitors ipilimumab, nivolumab, and pembrolizumab are used for a broad range of cancers and treatment is often accompanied by severe colitis. Results of this study revealed that differences in mode of action – ipilimumab inhibits CTLA-4 whereas nivolumab and pembrolizumab target PD-1 – accounted for a three-fold higher reporting rate of colitis linked to ipilimumab. Mapping chemical and biological data to adverse events reports suggested that ipilimumab may unblock the immune response earlier, leading to heightened T-cell activation.