Relevance of reactive metabolites and covalent binding to drug candidate selection

Idiosyncratic drug reactions (IDRs)represent a major risk for drug development. Current methods do not reliably predict IDR risk. There is strong evidence that most IDRs are immune-mediated. An adaptive immune response requires 2 signals: signal 1 represents recognition of drug-related antigens by T cell receptors presented in the context of HLA, and signal 2 represents upregulation of co-stimulatory molecules on antigen-presenting cells (APCs). There is circumstantial evidence that most IDRs are caused by reactive metabolites. Reactive metabolites have the potential to provide both signal 1 and signal 2. Covalent binding studies have been used to try to predict IDR risk, especially liver injury, but the results have been far from perfect with many safe drugs leading to high covalent binding. That begs the question of what reactive metabolite characteristics are associated with IDR risk. Likely characteristics associated with risk include dose, transporters that concentrate the drug in the liver, the enzyme that formed the reactive metabolite, the reactivity of the metabolite, and how it is presented to the immune system. For example, some drugs are bioactivated by myeloperoxidase, which is present in neutrophils and APCs. Not only is this associated with the risk of agranulocytosis, but it also can lead to activation of APCs and upregulation of signal 2. Unlike signal 1, which requires specific HLA molecules and T cell receptors, signal 2 is unlikely to be idiosyncratic. There is evidence that release of damage-associated molecular pattern molecules (DAMPs) and activation of APCs are better predictors of IDR risk.