Resolving 3D-domain swapping of human IL-10 prevents extensive multimerisation and increases yield

Heterologous expression platforms of biopharmaceutical proteins have been significantly improved over the last decade. Further improvement can be established by examining the intrinsic properties of proteins. Interleukin-10 (IL-10) is an anti-inflammatory cytokine with a short half-life that plays an important role in re-establishing immune homeostasis. This homodimeric protein of 36 kDa has significant therapeutic potential to treat inflammatory and autoimmune diseases. We have shown that the major production bottleneck of human IL-10 is not protein instability as previously suggested, but extensive multimerisation due to its intrinsic 3D domain swapping characteristic. Extensive multimerisation of human IL-10 could be visualised as granules in planta. On the other hand, mouse IL-10 hardly multimerised, which could be largely attributed to its glycosylation. By introducing a short glycine-serine-linker between the fourth and fifth alpha helix of human IL-10 a stable monomeric form of IL-10 (hIL-10mono) was created that no longer multimerised and increased yield up to 20-fold. However, hIL-10mono no longer had the ability to reduce pro-inflammatory cytokine secretion from lipopolysaccharide-stimulated macrophages. Forcing dimerisation restored biological activity. This was achieved by fusing human IL-10mono to the C-terminal end of constant domains 2 and 3 of human immunoglobulin A (Fca), a natural dimer. Stable dimeric forms of IL-10, like Fca-IL-10, may not only be a better format for improved production, but also a more suitable format for medical applications.