Spatial and Quantitative Optimization of Engineered Multi-Element Therapeutic Proteins

The current paradigm of drug development is to create single-function molecules that bind tightly to their targets and inhibit their function.  In contrast, natural processes, including defense against pathogens and disease, involve many elements – multiple protein domains, weak interactions that function cooperatively, geometrically optimized elements, and so on.  The spatial and quantitative aspects of such systems are tweaked in the course of evolution to create defenses that look quite different from the simple molecules designed by humans.  Our goal is to create complex therapeutic proteins and cells that mimic natural systems.  We have engineered proteins that deliver an activating hormone or cytokine, either erythropoietin or interferon alpha, to specific target cells.  The quantitative aspect is to mutate the cytokine and reduce its activity, so that cell binding is driven by an antibody element to which the cytokine is fused.  The spatial aspect is primarily to use a linker that allows simultaneous binding of the cytokine and antibody elements to their receptors, and also maximizes the rate of the cytokine binding after the antibody element has already bound.  We also developed course-grained molecular dynamics software to simulate the movement and binding of such proteins, with the goal of optimizing the engineering choices before making the proteins.

Erythropoietin (Epo) is a hormone, produced by the kidneys in response to low blood oxygen, that stimulates red blood cell production.  Commercial Epo (e.g. Procrit, Epogen) is used to treat anemia in kidney failure and cancer patients.  In 2004, sales were about $10B, but since then the use of Epo has dropped significantly due to clinical trials showing that treated patients suffered from clotting events leading to heart attacks, strokes, and deep vein thrombosis.  These effects are due to Epo action on cells other than red blood cell precursors, such as platelet precursors and blood vessel endothelial cells.  To avoid these side effects, we constructed a form of Epo that is fused to an antibody and targeted to late red blood cell precursors.  Upon treatment of mice, the engineered protein stimulates production of only red blood cells, while controls and commercial forms of Epo stimulate red blood cells and platelets to a similar extent.

Interferon alpha (IFNalpha) is a cytokine that represents the primary response to virus infection.  IFNalpha is used in treatment of hepatitis and certain cancers, but a major side effect is ‘flu-like symptoms’, so that long-term high-dose treatments are burdensome.  We generated a fusion protein consisting of a weakened IFNalpha fused to antibody V regions that bind to a tumor-specific surface marker.  Animal testing is in progress.

These experiments illustrate how protein engineering can be added to the repertoire of synthetic-biological tools to create novel functions that go beyond what is possible by manipulating transcriptional circuits.