Redirecting immune signaling with cytokine adaptors

Abstract

Cytokines are signaling molecules that coordinate complex immune processes and are frequently dysregulated in disease. While cytokine blockade has become a common therapeutic modality, cytokine agonism has had limited utility due to the widespread expression of cytokine receptors with pleiotropic effects. To overcome this limitation, we devise an approach to engineer molecular switches, termed cytokine adaptors, that transform one cytokine signal into an alternative signal with a different functional output. Endogenous cytokines act to nucleate the adaptors, converting the cytokine-adaptor complex into a surrogate agonist for a different cytokine pathway. In this way, cytokine adaptors, which have no intrinsic agonist activity, can function as conditional, context-dependent agonists. We develop cytokine adaptors that convert IL-10 or TGF-β into IL-2 receptor agonists to reverse T cell suppression. We also convert the pro-inflammatory cytokines IL-23 or IL-17 into immunosuppressive IL-10 receptor agonists. Thus, we show that cytokine adaptors can convert immunosuppressive cytokines into immunostimulatory cytokines, or vice versa. Unlike other methods of immune conversion that require cell engineering, cytokine adaptors are soluble molecules that leverage endogenous cues from the microenvironment to drive context-specific signaling.

Fig. 1. TGF-β→IL-2 cytokine adaptors compel IL-2 receptor signaling in the presence of TGF-β.

a Conceptual overview of an adaptor. b Schematic illustrating cytokine adaptors, which convert an inhibitory cytokine into a stimulatory cytokine by blocking the inhibitory cytokine from binding to its receptor and instead compelling dimerization of the stimulatory receptor. Model is based on IL-2R (PDB ID: 2B5I) and TGF-βR (PDB ID: 2PJY). c Model comparing natural TGF-β signaling through TGF-βRI and TGF-βRII (left) vs. signaling through TGF-β→IL-2 cytokine adaptors (right) which dimerize IL-2Rβ and gamma-c (γ_c). Adaptor molecules are comprised of scFv GC1008 linked to a VHH against IL-2Rβ (IL-2RβNb6) or a VHH against γ_c (γ_cNb6). d Cartoon representations of TGF-β Adaptor T.1 and Adaptor T.2. e, f Adaptors T.1 and T.2 signal through pSTAT5 in the presence of TGF-β. e Dose–response curves for phospho-STAT5 in YT-1 cells stimulated for 20 min with human IL-2 or TGF-β with equimolar Adaptor T.1 and Adaptor T.2. Data plotted as the mean of n = 2 technical replicates. Data are representative of N = 3 independent experiments (n = 2, N = 3). f Phospho-STAT5 E_max calculated from dose–response curves in YT-1 cells normalized to IL-2 E_max. Bar graphs represent mean, n = 2, N = 3. Source data are provided as a Source data file.

Fig. 2. Single-component cytokine adaptor T.3 converts TGF-β stimulation to IL-2 signaling in human primary T cells.

a Schematic illustrating rationale for designing single-component cytokine adaptors. In the two-component system (top), non-productive pairs of T.1/T.1 and T.2/T.2 are not able to dimerize IL-2Rβ and γ_c. To engineer a single-component adaptor (bottom), the two components are linked by a flexible linker, such that receptor dimerization is compelled only in the presence of TGF-β. b Several single-component cytokine adaptors were designed with varying orientations and linker lengths. Adaptors T.3, T.4, and T.5 contain N-terminal IL-2RβNb6 and linker lengths of 20aa, 30aa, or 10aa respectively, while Adaptor T.6 features a reversed orientation. c, d Single-component adaptors suppress TGF-β-induced pSMAD2/pSMAD3 signaling while stimulating pSTAT5 signaling in human CD4⁺ (c) and CD8⁺ (d) T cell blasts. Data is represented as mean fluorescence intensity of phospho-STAT5 (n = 2) and pSMAD2/pSMAD3 (n = 1) of T cells treated with 10 nM TGF-β +/- Adaptor T.3, T.4, T.5, or T.6. Bar graphs represent mean ± SD. e, f Single-component adaptor T.3 improves dose–response signaling in human primary T cells. Dose–response curves are shown for phospho-STAT5 in human CD4⁺ (e) and CD8⁺ (f) T cell blasts stimulated for 20 min with human IL-2, Adaptor T.3, TGF-β with equimolar Adaptor T.3, or TGF-β with equimolar Adaptor T.1 and Adaptor T.2. Data are plotted as mean, n = 2, N = 3. g Schematic illustrating the suppressive effects of TGF-β and stimulatory effects of IL-2 on T cell proliferation and cytotoxicity in the tumor microenvironment. h–j Adaptor T.3 increases CD8⁺ T cell proliferation (h), IFN-γ (i), and TNF-α (j) production in the presence of TGF-β. Bar graphs represent mean, n = 3, N = 3. * indicates p < 0.05, ** indicates p < 0.01 by one-way non-parametric ANOVA (Kruskal–Wallis test) with multiple comparisons. Bar graphs represent mean ± SD. Source data are provided as a Source data file.

Fig. 3. IL-10→IL-2 adaptors convert IL-10 stimulation into IL-2 signaling.

a Model comparing natural IL-10 signaling through IL-10Rα and IL-10Rβ (left, PDB ID: 6X93) vs. signaling through IL-10→IL-2 cytokine adaptors (right) which dimerize IL-2Rβ and γ_c. Adaptors are modeled using structures of IL-10 and scFv 9D7 (PDB ID: 1LK3) overlaid with IL-10 and IL-10Rα (PDB ID: 6X93), IL-2RβNb6 bound to IL-2Rβ (PDB ID: 7S2S) and γcNb6 bound to γc (PDB ID: 7S2R). b Cartoon representations of adaptors 10.1, 10.2, 10.3, and 10.4. c IL-10→IL-2 adaptors induce dose-dependent pSTAT5 signal similar to IL-2. Dose–response curves are shown for pSTAT5 in YT-1 cell line stimulated for 20 min with human IL-2 or IL-10 with equimolar Adaptor 10.1 and 10.2 or Adaptor 10.3 and 10.4. Data are plotted as mean, n = 2, N = 3. d Normalized pSTAT5 response in YT-1 cells stimulated with the indicated cytokines or adaptors, normalized as a percentage of maximum IL-2 pSTAT5. Data are plotted as mean, n = 2. e Adaptors 10.1 and 10.2 suppress IL-10-induced pSTAT3 and promote IL-2 like pSTAT5 signaling. Dose–response curves for pSTAT5 and pSTAT3 are shown in human CD4⁺ T cell blasts stimulated with 10 nM IL-10 and increasing concentrations of equimolar Adaptor 10.1 and 10.2. n = 2 replicates per condition. Source data are provided as a Source data file.

Fig. 4. IL-23→IL-10 cytokine adaptors mimic effects of IL-10.

a Model comparing natural IL-23 signaling through IL-23R and IL-12Rbβ1 (left, PDB ID: 6XDQ) vs. signaling through IL-23→IL-10 cytokine adaptors (right) which dimerize IL-10Rα and IL-10Rβ. Model of IL-23→IL-10 adaptors is comprised of IL-23p19 binder VHH 37D5 and IL-23p40 binder VHH 22E11 and IL-23p19 (PBD ID: 4GRW) overlaid with structures of IL-10 receptor complex (PDB ID: 6X93). b Cartoon representation of IL-23→IL-10 cytokine adaptor design. c, d IL-23→IL-10 adaptors induce a pSTAT3 signal in THP-1 cells similar to IL-10. c Dose–response curves for phospho-STAT3 in THP-1 cell line stimulated for 20 min with human IL-10, a supermonomeric version of IL-10, or IL-23 with equimolar IL-23→IL-10 cytokine adaptors. Data plotted as mean phospho-STAT3 mean fluorescence intensity, n = 2, N = 3. d Normalized pSTAT3 response in THP-1 cells stimulated with 10 nM WT IL-10, a supermonomeric variant of IL-10 (M), a monomeric dominant negative mutant of IL-10 (DN), IL-23, or IL-23 with the IL-23 → IL-10 adaptors. n = 2, N = 3. Data are plotted as mean ± SD. e–g IL-23→IL-10 adaptors suppress LPS-induced production of inflammatory cytokines. TNF-α (e), IL-6 (f), and IL-1β (g) quantified by ELISA in supernatants from human PBMCs stimulated for 24 h with LPS and 10 nM IL-10, IL-23, or IL-23 → IL-10 adaptors. n = 3 replicates per condition, N = 3. Bar graphs represent mean ± SD. Source data are provided as a Source data file.

Fig. 5. IL-17→IL-10 cytokine adaptors mimic effects of IL-10.

a Model comparing endogenous IL-17A signaling through IL-17RA and IL-17RC (left, PDB ID: 7UWN) vs. signaling through IL-17→IL-10 cytokine adaptors (right) which dimerize IL-10Rα and IL-10Rβ. Model of IL-17→IL-10 adaptors is comprised of IL-17A binders CAT2200 (PBD ID: 2VXS) and netakimab fragment VHH-76 (PDB ID: 8B7W) overlaid with structures of IL-10 receptor complex (PDB ID: 6X93). b Cartoon representations of IL-17→IL-10 cytokine adaptor designs with 10 amino acid (aa) linker between modules. c, d IL-17→IL-10 adaptors induce a dose-dependent pSTAT3 signal in THP-1 cells. c Dose–response curves for phospho-STAT3 in THP-1 cell line stimulated for 20 min with human IL-10, a supermonomeric version of IL-10 or IL-17A with equimolar IL-17→IL-10 cytokine adaptors. Data represented as mean, n = 2, N = 3. d Normalized pSTAT3 response in THP-1 cells stimulated with 10 nM WT IL-10, human IL-10, a supermonomeric version of IL-10 (M), IL-17A, or IL-17A with the IL-17→IL-10 adaptors. Data are plotted as mean, n = 2, N = 3. e–g IL-17→IL-10 adaptors suppress LPS-induced production of inflammatory cytokines. TNF-α (e), IL-6 (f), and IL-1β (g) quantified by ELISA in supernatants from human PBMCs stimulated with 24 h with LPS and 10 nM IL-10, IL-17, or IL-17→IL-10 adaptors. Bar graphs represent mean, n = 2. Source data are provided as a Source data file.