Re-targeting of anthrax toxin binding for immunomodulation and targeted cancer therapy

Abstract

The intracellular delivery of cytotoxic proteins is a longstanding goal in drug development. This is challenging due to biological membranes, which prevent facile entry of macromolecules into the cellular cytosol. Anthrax, derived from B. anthracis, is a protein delivery platform. Recent work has established anthrax as a non-toxic and reliable method to deliver a variety of proteins and other molecules to the cytosol of cells in a receptor-directed manner. However, native anthrax receptors are ubiquitously expressed membrane proteins TEM8 and CMG2, limiting the therapeutic application of wild-type anthrax. Moreover, unmodified anthrax is immunogenic, which limits the efficacy and number of doses which can be administered. In this thesis, anthrax was re-targeted to cells of therapeutic interest.

In particular, the pore-forming agent of anthrax, protective antigen (PA), was fused to single-chain variable fragments targeting two pancreatic cancer cell receptors, human endothelial growth factor receptor and carcinoembryonic antigen. This design enabled pancreatic cancer cell-specific delivery of two highly toxic protein cargoes, including a Ras protease which rapidly degrades cytosolic Ras protein. The mutant protective antigen-scFv provides a facile and generalizable strategy for delivery of cytosolic delivery of proteins to any number of membrane receptors which can be targeted by known scFvs and, by extension, known antibodies. To mitigate anthrax's inherent immunogenicity, anthrax was re-targeted to mouse erythrocytes. Recent work has established that targeting of antigens to erythrocytes can lead to decreased antigen-specific inflammatory responses.

To discover a reliable method to direct a molecule, such as protective antigen from anthrax toxin, to erythrocytes, a synthetic peptide library was selected in mice to identify a D-peptide, DQLR, which binds preferentially to mouse erythrocytes in vivo. When administered to immunocompetent mice, a DQLR-PA conjugate led to significantly decreased anti-PA antibodies. Similarly, a DQLR-peptide antigen conjugate led to decreased antigen-specific inflammatory responses and antigen-specific T cells, indicating antigen-specific tolerance is induced by the DQLR-mediated association of antigens to mouse erythrocytes in vivo.