Engineered Nanobody Drug Tames Inflammation by Rewiring IL-10 Receptor Geometry

Interleukin-10 has long been one of immunology’s most tantalizing and most frustrating molecules. The cytokine is a master regulator of immune tolerance, damping down the inflammatory firepower of monocytes, macrophages and dendritic cells, and loss-of-function mutations in IL-10 or its receptors cause devastating early-onset inflammatory bowel disease. Yet every attempt to turn IL-10 into a drug for chronic inflammation has stumbled over the same paradox: at the doses needed to calm myeloid cells, the molecule also activates T cells, drives interferon-gamma-dependent myelopoiesis and depletes platelets and red blood cells. Early clinical trials in Crohn’s disease and psoriasis revealed a narrow therapeutic window punctuated by dose-limiting anaemia and thrombocytopaenia. Now a team of researchers reports a way out of this dilemma, not by modifying the cytokine itself but by rebuilding the geometry of its receptor complex with an engineered antibody scaffold.

Writing in Nature, the group describes a myeloid-selective interleukin-10 receptor agonist, or msIL-10R agonist, a tetravalent nanobody-based molecule that reproduces IL-10’s potent suppression of inflammatory myeloid cells while leaving T cells untouched. The key insight came from a careful dissection of how different immune cells transmit the IL-10 signal. IL-10 normally binds IL-10Rα and IL-10Rβ to form a tetrameric receptor complex constitutively associated with the Janus kinases JAK1 and TYK2. When the cytokine engages the receptor, JAK1 must be activated through transphosphorylation, and the team found that cells can accomplish this through two parallel routes: TYK2-to-JAK1 or JAK1-to-JAK1. Crucially, the relative reliance on these routes differs dramatically between cell types.

To map these routes, the researchers used a selective allosteric JAK1 inhibitor called VVD-118313, which blocks JAK1 transphosphorylation without touching the kinase’s active site, alongside the TYK2 inhibitor zasocitinib and CRISPR-based gene knockout. In mouse bone-marrow-derived macrophages, blocking JAK1-to-JAK1 transphosphorylation only partially reduced IL-10-induced STAT3 phosphorylation and did not abolish the cytokine’s functional suppression of tumour necrosis factor. Only when TYK2 was also removed did signalling collapse completely, showing that at least one transphosphorylation route must remain intact. Human monocytes behaved similarly: TYK2 inhibition or knockout dented STAT3 phosphorylation but left the suppression of LPS-induced TNF and IL-6 fully intact.

T cells told a strikingly different story. In CD8-positive T cells, both pharmacological TYK2 inhibition and CRISPR knockout diminished IL-10-induced STAT3 phosphorylation and completely eliminated the cytokine’s augmentation of granzyme B and interferon-gamma production. T cells, in other words, depend strictly on TYK2 and on maximal pathway activation, whereas myeloid cells tolerate a substantially attenuated signal. The team confirmed this threshold logic with an engineered IL-10 monomer that engages only a single IL-10Rα–IL-10Rβ heterodimer, restricting signalling to one TYK2–JAK1 route. The monomer drove roughly half of maximal STAT3 phosphorylation in monocytes yet retained meaningful TNF suppression, while failing entirely to activate STAT3 or cytotoxic effector programs in T cells. Attenuation, it seemed, could be a feature rather than a bug.

Attenuated cytokines alone, however, make poor drugs. Previous IL-10 monomers and affinity-tuned muteins suffered from short half-lives, thermal instability, residual T and B cell activation and vulnerability to naturally occurring anti-IL-10 autoantibodies, which are common in inflammatory bowel disease. The team instead turned to single-domain antibodies. Alpacas were immunized with the extracellular domains of human IL-10Rα and IL-10Rβ, and phage-display libraries built from their B cells yielded diverse families of receptor-binding VHH domains. These were assembled into dimeric and tetravalent constructs pairing anti-IL-10Rα with anti-IL-10Rβ nanobodies, and screened in STAT3 reporter cells and primary human cells. Of 882 combinations tested, 182 showed reporter activity and 21 achieved more than 90 percent inhibition of TNF and more than 85 percent inhibition of IL-6 in LPS-stimulated blood cells.

Most of those hits, however, still recapitulated IL-10’s unwanted T cell stimulation. One construct stood apart. Designated the msIL-10R agonist, it pairs an anti-IL-10Rα nanobody called E07 with an anti-IL-10Rβ nanobody called A05 in each arm of an Fc scaffold, creating a tetravalent molecule with picomolar potency. It inhibited LPS-induced TNF with a half-maximal inhibitory concentration of 2.1 picomolar and IL-6 at 6.9 picomolar in human blood cells, while eliciting less than 10 percent of maximal STAT3 phosphorylation in CD4 and CD8 T cells, no induction of interferon-gamma or granzyme B, and no B cell proliferation or plasmablast differentiation. RNA sequencing confirmed that the molecule reproduced IL-10’s anti-inflammatory transcriptional program in resting and LPS-stimulated blood cells, including suppression of CSF3, IL6 and PTGES, yet induced no significantly differentially expressed genes in activated T cells.

Cryo-electron microscopy revealed why geometry matters so much. Using a fiducial Fab to enlarge the small receptor–nanobody complex, the team resolved the IL-10Rα–IL-10Rβ–E07-A05 assembly at 3.14 angstroms and a second IL-10Rα–E07 complex at 2.40 angstroms. The nanobody A05 binds an epitope on IL-10Rβ that partially overlaps the cytokine-binding surface but induces a local rearrangement of the Asn138–Trp143 loop, while E07 recognizes a broad surface spanning the junction of IL-10Rα’s two fibronectin type III domains, on the opposite side of the receptor from the IL-10 binding site. Notably, disease-associated IL10RA mutations such as Y64C and R117C, which cripple native IL-10 signalling and underlie very-early-onset inflammatory bowel disease, left the msIL-10R agonist’s activity fully intact, because the molecule engages epitopes spatially distinct from the cytokine interface.

Functional dissection showed that the tetravalent architecture is not negotiable. Disrupting either receptor-binding arm abolished STAT3 phosphorylation and cytokine suppression, maximal activity required simultaneous cross-linking of two IL-10Rα molecules with one or two IL-10Rβ molecules, and even subtle geometric changes were consequential. Shortening or lengthening the glycine-serine linker between E07 and A05 reduced activity, inverting the tandem nanobody order to A05-E07 abolished signalling entirely, and a symmetric A05-A05/E07-E07 arrangement boosted myeloid potency but restored roughly 20 percent of native IL-10’s interferon-gamma response in T cells. Signalling output, the authors conclude, is governed not simply by receptor occupancy but by the precise topology of receptor cross-linking, which forces the complex into a geometry that permits only TYK2-to-JAK1 transphosphorylation.

That strict TYK2 dependence, combined with receptor abundance, explains the selectivity. Myeloid cells express two- to eighteen-fold more surface IL-10Rα and IL-10Rβ than T, B or natural killer cells, providing the avidity needed to cluster the tetravalent agonist, whereas lymphocytes fall below the threshold. When the researchers forced T cells to overexpress IL-10Rα, and especially both receptor chains, STAT3 phosphorylation and granzyme B production returned, confirming receptor copy number as the critical determinant. The logic even proved portable: bispecific nanobodies pairing IL-10Rα with the lymphocyte-enriched IL-2Rγ chain redirected signalling toward T and NK cells, producing T/NK-selective agonists with minimal monocyte activity, again exquisitely dependent on geometry.

In vivo, the molecule delivered. In humanized IL10RA/IL10RB knock-in mice, it suppressed systemic TNF, IL-6 and IL-12 after LPS challenge as effectively as IL-10-Fc, reduced ear swelling and lymph node neutrophil infiltration in a delayed-type hypersensitivity model, and improved colon weight-to-length ratios, splenic neutrophil expansion and inflammatory gene expression in DSS-induced colitis. In cynomolgus monkeys, a humanized version showed dose-proportional exposure, sustained monocyte receptor occupancy and a terminal half-life of roughly eight days, with durable TNF suppression for more than three weeks after a single dose. Unlike IL-10-Fc, which elevated IL-18 and caused anaemia and thrombocytopaenia, the agonist produced no IL-18 elevation or haematological abnormalities at doses up to 100 milligrams per kilogram. The work establishes a geometry-based strategy for cell-type-selective cytokine engineering and positions the msIL-10R agonist as a promising candidate for inflammatory bowel disease and other myeloid-driven conditions, while offering a generalizable blueprint for precision cytokine therapeutics across immunology and oncology.

Subject of Research: Engineering a myeloid-selective IL-10 receptor agonist through TYK2-dependent signalling and receptor geometry

Article Title: Engineered TYK2 dependence enables myeloid-selective IL-10 receptor agonism

Article References: Engineered TYK2 dependence enables myeloid-selective IL-10 receptor agonism. (n.d.). https://doi.org/10.1038/s41586-026-11176-1

Image Credits: AI Generated

DOI: 10.1038/s41586-026-11176-1

Keywords: interleukin-10, IL-10 receptor, TYK2, JAK1, nanobody, myeloid cells, T cells, STAT3, cryo-electron microscopy, inflammatory bowel disease, cytokine agonism, immunotherapy