A Comprehensive Mathematical Model of Avidity in Cytokine Signaling

Abstract

Cytokine sensitivity varies substantially across cell types and cellular states, in part through differences in the abundance of individual receptor subunits. Yet for multicomponent receptors, the quantitative relationship between receptor abundance, binding affinity, and functional potency remains poorly defined. Here, we derive closed-form expressions for EC50 in multivalent ligand-receptor systems at cell surfaces. Unlike monovalent models, these equations predict that potency depends on both binding constants and receptor abundance, with individual receptor subunits playing asymmetric roles in controlling maximal complex formation and cellular sensitivity. We validate these predictions using quantitative antibody-binding, cytokine-binding, and signaling data, and extend the equilibrium framework to steady-state signaling conditions in which downstream kinetic processes can impose additional limits on functional potency. We then derive a regression-compatible formulation and test its predicted ligand-receptor-response relationships across an in vivo murine cytokine perturbation atlas, human spatial transcriptomic data, and an independent cohort of 510 patients with lung adenocarcinoma. Multivariable cytokine relationships were conserved across human datasets and associated with clinical outcome. In addition, quantitative proteomics from primary human immune cells supported receptor mRNA abundance as an informative but imperfect proxy for protein abundance. Together, this framework connects multivalent receptor biophysics with cell-type-specific signaling across molecular, cellular, tissue, and patient scales.