Trioxacarcin-based Antibody-Drug Conjugates: Optimization of payload structure and a triazene linker system design

Abstract

The trioxacarcins are a set of rigid, pentacyclic natural products first isolated from Streptomyces bottropensis DO-45 in 1981. Their exceptional cytotoxicity inspired a phase 1 clinical trial only a few years later, which was halted abruptly due to unforeseen cardiotoxicity, resulting in a participant’s passing. Forgotten for decades, interest in trioxacarcins resurged more recently with the advent of contemporary, targeted cancer therapies, including Antibody-Drug Conjugates (ADCs), which promise to deliver the toxic payload directly to the tumor, thereby improving the therapeutic index. Following the initial disclosure of the total synthesis of Trioxacarcin A and related natural products back in 2013, the Myers group has been hard at work developing simplified, scalable pharmacophores that maintained the impressive potency of the initial bacterial isolates, and incorporated those into ADCs. Unfortunately, the very functionality responsible for its activity – the unique spiro-epoxide capable of alkylating guanine nucleobases in DNA – proved to be incompatible with the ADC platform, reacting with the nucleophilic amino acid side chains of the monoclonal antibody instead. Additional efforts resulted in the development of a bromohydrin-based prodrug strategy, however these drug-linkers lacked the requisite activity. This thesis sets out to understand the underpinning reasons for the lack of ADC activity in several complementary ways. In Chapter 2, I present the final attempt at synthesizing trioxacarcin ADCs with an intact spiro-epoxide, using chemical methods to install a functional handle near the reactive functional group, in the hope of shielding it from the monoclonal antibody. Chapter 3 identifies the bromohydrin reversion kinetics as the likely culprit of the lack of activity. Using a hypothesis-driven approach, we synthesize a panel of new trioxacarcin analogs with the goal of improving the efficiency of the spiro-epoxide reformation. We then identify a payload with the reversion kinetics improved 7-fold over the initial lead compound, and prove that the ADC activity is directly related to the rate of this process. Unfortunately, the improved reversion kinetics come at a cost of lower warhead potency, leading to equipotent ADCs. In Chapter 4 I describe the design and synthesis of a novel drug-linker system inspired by the glioblastoma drug temolozomide. It incorporates a highly reactive triazene motif, which fragments in aqueous buffers to reveal a diazonium cation that immediately reforms the parent spiro-epoxide. By synthesizing different drug-linkers I show how the fate of the warhead is directly related to the substitution pattern of the triazene. The new constructs exhibit vastly superior epoxide reformation kinetics, with full conversion within 20 minutes versus multiple hours observed for the bromohydrin systems.