Receptor dimerization is a fundamental mechanism by which most cytokines and growth factors activate Type-I transmembrane receptors. Although previous studies have shown that ligand-induced topological changes in the extracellular domains (ECDs) of dimeric receptors can affect signaling output, the physiological relevance is not well understood. This is a difficult problem to study because an engineered ligand system does not exist that would enable a systematic exploration of the relationship between ligand-receptor dimer geometry and signaling. This question, as well as the capability of exploiting these structure-activity relationships for drug discovery, are important because most cytokines exert pleiotropic effects that limit their therapeutic utility. New approaches are needed to modulate cytokine signaling and identify clinically efficacious variants. By contrast, small-molecule ligands for G protein–coupled receptors have been successfully discovered through medicinal chemistry and used to induce conformational changes that lead to physiologically relevant biased signaling outputs. A comparable approach for dimeric receptors could open a path to new pharmacological parameters for cytokines and growth factors.

In order to better understand how the extracellular structure of cytokine-receptor complexes affects downstream signaling events, we developed an engineered ligand system to precisely control the orientation and proximity of dimeric receptor complexes that would enable the measurement of structure-activity relationships between receptor dimer geometry, signaling, and function. We applied this approach to design geometrically controlled ligands to the erythropoietin receptor (EpoR) system, a well-characterized dimeric cytokine receptor system.

We used the DARPin (designed ankyrin repeat protein) scaffold because of its modular nature. We isolated a high-affinity DARPin to EpoR using yeast display and in vitro evolution and determined the crystal structure of the DARPin/EpoR complex. We then converted these monomeric DARPin binding modules into C2 symmetric homodimeric agonists by incorporating in silico designed dimerization interfaces. This rigidly connected dimeric DARPin scaffold then enabled us to design a series of extended ligands through sequential insertion of ankyrin repeat “spacers” to systematically control the relative orientation of the ECDs in the dimeric complex. The “angle” series varied the scissor angle between the two ECDs, whereas the “distance” series varied their relative proximity. The designed DARPin ligands were validated by means of x-ray crystallography for representative complexes. The systematic variation of angular and distance parameters generated a range of full, biased, and partial agonism of EpoR signaling in the human erythroid cell line UT7/EPO, as shown with flow-cytometry and immunoblotting for phosphorylated downstream effectors. In general, increasing the angle or distance between the receptor ECDs resulted in a progressive partial agonism, as measured with changes in maximum response achieved (E_max) and median effective concentration (EC₅₀). Biased signal transducer and activator of transcription (STAT) activation was elicited by some of the surrogate DARPin ligands. We also evaluated the effects of these DARPin agonists on differentiation and proliferation of hematopoietic stem and progenitor cells (HSPCs) that were maturing into the erythroid lineage. The partial agonists displayed stage-selective effects on HSPCs, whereas the biased agonists more selectively promoted signaling at either the early or late stages …