A synthesis enabled understanding of Amphotericin B leading to derivatives with improved therapeutic indices

Systemic fungal infections represent a significant contributor to human morbidity and mortality. To lower the impact of invasive fungal pathogens on the global population, more progress must be made in the areas of diagnostic tests and safe and effective new therapies. Although Amphotericin B (AmB) has served as the last line of defense against invasive fungal infections without significant pathogen resistance, its clinical application is restricted by severe dose-limiting toxicity. Efforts to improve AmB’s therapeutic index (a comparison between the dose that causes the desired therapeutic effect to the dose that causes toxicity) have not been successful. This lack of progress is almost completely attributed to an incorrect understanding of AmB’s primary mechanism of action. For more than 5 decades AmB has been perceived to primarily exist and operate as self assembled an ion channel complex—exerting cytotoxicity through the efflux of intracellular ions leading to the disruption of critical electrochemical gradients. Although a beautiful and rare phenomenon for a small molecule, the ion channel model has masked AmB’s underlying mechanism of action. In this body of work we will discuss how, through the systematic synthesis of single functional group deficient derivatives of AmB, coupled with state-of-the-art biophysical and biological experiments we have been able to fully elucidate AmB’s mechanism of action. We discuss how AmB predominantly exists as an extramembranous sterol sponge that primarily kills yeast and human cells by binding and extracting sterols in a mycosamine dependent fashion. Additionally, the C2’-OH and C3’-NH3+ are critical residues on the mycosamine appendage, which are responsible for stabilizing a conformation that allows for the binding of both ergosterol and cholesterol. When either of these residues is deleted, AmB can still bind ergosterol but can no longer bind cholesterol. This shift in sterol binding directly correlates to a substantial loss of toxicity. These results suggest that the C2’-OH and C3’-NH3+ do not directly bind sterols but are sites of allosteric modification. Armed with an accurate macroscopic and atomistic understanding of AmB, we were able to rationally guide the development of a novel derivative with an increased therapeutic index