Synthetic interleukin 22 (IL-22) signaling reveals biological activity of homodimeric IL-10 receptor 2 and functional cross-talk with the IL-6 receptor gp130

Cytokine signaling is transmitted by cell-surface receptors that function as biological switches controlling mainly immune-related processes. Recently, we have designed synthetic cytokine receptors (SyCyRs) consisting of GFP and mCherry nanobodies fused to transmembrane and intracellular domains of cytokine receptors that phenocopy cytokine signaling induced by non-physiological homo- and heterodimeric GFP-mCherry ligands. Interleukin 22 (IL-22) signals via both IL-22 receptor a 1 (IL-22R a 1) and the common IL-10R2, belongs to the IL-10 cytokine family, and is critically involved in tissue regeneration. Here, IL-22 SyCyRs phenocopied native IL-22 signal transduction, indicated by induction of cytokine-dependent cellular proliferation, signal transduction, and transcriptome analysis. Whereas homodimeric IL-22R a 1 SyCyRs failed to activate signaling, homodimerization of the second IL-22 signaling chain, SyCyR (IL-10R2), which previously was considered not to induce signal transduction, led to induction of signal transduction. Interest-ingly, the SyCyR(IL-10R2) and SyCyR(IL-22R a 1) constructs could form functional heterodimeric synthetic receptor chain SyCyR(gp130). homodimeric receptor broad receptor cross-talk of IL-22R IL-20R2

Cytokine signaling is transmitted by cell-surface receptors that function as biological switches controlling mainly immunerelated processes. Recently, we have designed synthetic cytokine receptors (SyCyRs) consisting of GFP and mCherry nanobodies fused to transmembrane and intracellular domains of cytokine receptors that phenocopy cytokine signaling induced by nonphysiological homo-and heterodimeric GFP-mCherry ligands. Interleukin 22 (IL-22) signals via both IL-22 receptor a1 (IL-22Ra1) and the common IL-10R2, belongs to the IL-10 cytokine family, and is critically involved in tissue regeneration. Here, IL-22 SyCyRs phenocopied native IL-22 signal transduction, indicated by induction of cytokine-dependent cellular proliferation, signal transduction, and transcriptome analysis. Whereas homodimeric IL-22Ra1 SyCyRs failed to activate signaling, homodimerization of the second IL-22 signaling chain, SyCyR (IL-10R2), which previously was considered not to induce signal transduction, led to induction of signal transduction. Interestingly, the SyCyR(IL-10R2) and SyCyR(IL-22Ra1) constructs could form functional heterodimeric receptor signaling complexes with the synthetic IL-6 receptor chain SyCyR(gp130). In summary, we have demonstrated that IL-22 signaling can be phenocopied by synthetic cytokine receptors, identified a functional IL-10R2 homodimeric receptor complex, and uncovered broad receptor cross-talk of IL-22Ra1 and IL-20R2 with gp130.
Cytokines control immune responses but are also involved in homeostatic processes, such as development, differentiation, growth, and regeneration. Signal transduction of cytokines is executed by natural biological switches, and among many other functions, it controls immunity-related processes (1). Cytokines switch transmembrane receptors from the off-state into the on-state via receptor dimerization or multimerization. The on-state might be interrupted by negative feedback mechanisms or depletion of the cytokine and cytokine receptor. Recently, we have designed synthetic cytokine receptors (SyCyRs), which phenocopy IL-6 and IL-23 signaling (2). SyC-yRs are based on nanobodies specifically recognizing GFP and mCherry (3,4) fused to the transmembrane and intracellular domains of the receptor of interest. The nanobodies serve as extracellular sensors for homo-and heteromeric GFP-mCherry fusion proteins, which induce receptor dimerization (5). A nanobody or VHH domain consists of the N-terminal variable domain of Camelidae heavy chain antibody, which is sufficient for antigen binding (6). Synthetic cytokine receptors might become important tools for immunotherapeutic applications (7), with chimeric antigen receptor (CAR) T-cell therapy being the first example that has been approved as gene therapy for the treatment of severe cases of acute lymphatic leukemia (8).
Moreover, synthetic cytokine biology can decipher the potential of cytokine receptor cross-talk. In a reductionistic view, a cytokine binds only to its corresponding cytokine receptor complex, which is composed either of receptor homo-or heterodimers. This simple view has been challenged for many cytokines and cytokine receptors that have multiple binding partners. For example, the signal receptor complex of IL-6 consists of two gp130 receptor chains, but gp130 is also the only receptor for IL-11 and the co-receptor for IL-27, CNTF, CT-1, LIF, and OSM. On the other hand, IL-35 from the IL-12-type cytokine family was proposed to activate a variety of different receptor complexes, including gp130 homodimers, IL-12Rb2 homodimers, and IL-12Rb2/WSX-1 and gp130/IL-12Rb2 heterodimers (9,10). Using chimeric cytokine receptors, we have shown that gp130 can form biologically active complexes with IL-23R, IL-12Rb2, and IL-12Rb1 of the closely related IL-12type cytokine family (11). The Interleukin 10 family consists of six members, with IL-10, IL-22, and IL-26 belonging to the IL-10 family and IL-24, IL-20, and IL-19 belonging to the IL-20 subfamily (12,13). There are three more distantly related cytokines that are sometimes classified as IL-10 family members or as type III interferons (IFNs): IL-28A (IFN-l2), IL-28B (IFN-l3), and IL-20 (IFN-l1) (14). IL-22 signals via the IL-10R2 and IL-22Ra1 and mainly by activation of Jak-mediated STAT3 phosphorylation and, albeit to a lesser extent, also STAT1, STAT5, and ERK. Jak1 and Tyk2 are preferentially used by IL-10R2 and IL-22Ra1 (15,16). IL-22 is predominantly produced by T-cell subsets and group 3 innate lymphoid cells (ILC3s) (17). In the intestinal epithelium, IL-22 is mainly responsible for immune homeostasis as well as wound healing (18) and plays important roles in the pathogenesis of many intestinal diseases. The majority of preclinical studies support a protective role for IL-22 against various pathogens, including bacteria, yeasts, viruses, and parasites (12). Therefore, mimicking IL-22 signal transduction in synthetic biology is of importance for future therapeutic applications. Here, we generated and characterized synthetic cytokine receptors for IL-22, revealed a novel role of the IL-10R2 in signal transduction, and found a functional cross-talk with the IL-6 receptor chain gp130.
Further, we analyzed the mRNA expression by gene array analysis of Ba/F3/gp130/IL-10R2/IL-22Ra1 cells stimulated with IL-22 and Ba/F3/gp130/C VHH IL-10R2/G VHH IL-22Ra1 cells stimulated with GFP-mCherry, indicating a high overlap of gene regulation. In Fig. 3A, all conditions were compared between the different cell lines in one scattered blot, which revealed a high degree of overlap. In Fig. 3 (B and C), we specifically compared unstimulated and stimulated conditions of one cell line and observed an activation of gene transcription upon stimulation that was stronger for IL-22 compared with the GFP-mCherry stimulated cell line. However, among the regulated genes are, in both conditions, typical STAT3 target genes, including SOCS3, Pim-1, and Myc (Fig. 3D). mRNA level of SOCS3, Pim-1, and Myc was verified by qPCR (Fig. 4), supporting our data obtained from the gene array analysis. In summary, our data revealed a high degree of overlap between the signaling induced by the synthetic ligand GFP-mCherry and the natural cytokine IL-22.
SyCyRs for IL-22Ra1 fail to form biologically active homodimers Next, we used Ba/F3/gp130/G VHH IL-22Ra1 cells (Fig. S2A) to verify that homodimers of the synthetic IL-22Ra1 are not biologically active as suggested in Fig. 1C. Cell-surface expression of G VHH IL-22Ra1 was verified by flow cytometry (Fig.  S2B). Dimeric GFP-Fc was not able to induce cellular proliferation of Ba/F3/gp130 and Ba/F3/gp130/G VHH IL-22Ra1 cells (Fig. S2C). Also, monomeric and dimeric GFP-Fc did not induce STAT3 and ERK1/2 phosphorylation (Fig. S2D). As a control, Hyper-IL-6 induced cellular proliferation and STAT3 and ERK1/2 phosphorylation (Fig. S2, C and D). We also performed this experiment with Ba/F3/gp130/C VHH IL-22Ra1 cells and dimeric mCherry and obtained comparable results (Fig. S3,  A and B). Interestingly, the overall assembly of G VHH IL-22Ra1 was correct because we showed binding of GFP-Fc to Ba/F3/ gp130/G VHH IL-22Ra1 but not to Ba/F3/gp130 cells by flow cytometry (Fig. S3C). In conclusion, SyCyRs for IL-22Ra1 fail to form biologically active homodimers. This finding was surprising because the intracellular domain of the long-chain cytokine receptor IL-22Ra1 is 346 amino acid residues long, and apart from binding sites for Jak1, it also contains multiple initiation sites for signal transduction (e.g. STATs) (25).

SyCyRs for IL-10R2 form biologically active homodimers
Receptors with short ICDs including IL-10R2 often bind their ligands with lower affinity, pair with Tyk2 or Jak2 (27), and are generally considered to minimally contribute to STAT recruitment and activation (28,29). Only in combination with long-chain receptors, such as IL-22Ra1, do this kind of receptors contribute to activation of signal transduction. On the other hand, results presented in Figs. 1 and 2 suggest that homodimers of IL-10R2 induce signal transduction that is highly similar to signaling induced by IL-10R2/IL-22Ra1 heterodimers. In line with this, Kotenko et al. (27) already showed that a chimeric homodimer of human IL-10R2 is able to phosphorylate STAT1. Therefore, we generated Ba/F3/gp130/ C VHH IL-10R2 cells (Fig. 5A). Cell-surface expression of C VHH IL-10R2 was verified by flow cytometry (Fig. S4A). The proliferation of Ba/F3/gp130/C VHH IL-10R2 cells depended on the concentration of synthetic mCherry-Fc fusion proteins, reaching the half-maximal proliferation at 41.74 ng/ml (Fig.  5B). mCherry-Fc-induced proliferation of Ba/F3/gp130/ C VHH IL-10R2 cells was inhibited by P6 (Fig. 5C). Western blotting of Ba/F3/gp130/C VHH IL-10R2 cells showed that mCherry-Fc induced STAT3, ERK1/2, Jak1, Jak2, and Tyk2 phosphorylation, whereas monomeric mCherry was not able to induce signal transduction (Fig. 5D). P6 selectively inhibited phosphorylation of Jak1 but not of Jak2 and Tyk2 and resulted in STAT3 suppression and ERK1/2 phosphorylation (Fig. 5D). The inhibition of Jak2 and Tyk2 phosphorylation was achieved using higher concentrations of 100 mM P6 (Fig. S1C). The IL-10R2 was also able to phosphorylate STAT3 when expressed with a GFP VHH , resulting in Ba/F3/gp130/G VHH IL-10R2 cells (Fig. S4B). The finding that C VHH IL-10R2 homodimerization induced activation of Jak1, Jak2, and Tyk2 was supported by coimmunoprecipitation experiments using transiently transfected HEK293T cells. Here, precipitation of Jak1, Jak2, and Tyk2 resulted in co-immunoprecipitation of G VHH IL-10R2, suggesting that all three tyrosine kinases are physically interacting with the intracellular domain of G VHH IL-10R2 (Fig. 5E).
Next, we analyzed the mRNA expression by gene array analysis of Ba/F3/gp130/C VHH IL-10R2/G VHH IL-22Ra1 cells stimulated either with GFP-mCherry or mCherry-Fc, indicating a high overlap of gene regulation (Fig. 6A). In Fig. 6 (B and C), we specifically compared unstimulated and stimulated conditions either as homo-or heterodimer of the cell line and observed an activation of gene transcription upon stimulation that was comparable between the synthetic cytokine stimulations. However, among the regulated genes are, in both conditions, typical STAT3 target genes, including SOCS3, Pim-1, and Myc ( Fig.  6D). mRNA level of SOCS3, Pim-1, and Myc was verified by qPCR ( Fig. 6E), supporting our data obtained from the gene array analysis. This indicates that the IL-10R2 homodimer can induce nearly the same signal transduction as the IL-10R2/IL-22Ra1 heterodimer.
Western blotting data show one representative experiment of three.
not facilitate cellular proliferation with 100 ng/ml mCherry-Fc (Fig. 7B). At this concentration of mCherry-Fc, homodimeric C VHH IL-10R2(D330) but not C VHH IL-10R2(D310) expression of Pim1 mRNA was induced as quantified by qPCR (Fig. 7C). The EC 50 of Ba/F3/gp130/C VHH IL-10R2(D330) cells was 54.78 ng/ml mCherry-Fc and comparable with the EC 50 of Ba/F3/ gp130/C VHH IL-10R2 cells of 51.42 ng/ml mCherry-Fc in this experiment. Interestingly, also Ba/F3/gp130/C VHH IL-10R2 (D310) slightly proliferated in the presence of mCherry-Fc, albeit at a much higher concentration and did not reach maximal proliferation at 1000 ng/ml mCherry-Fc (Fig. 7D). Next, we analyzed STAT3, ERK1/2, Jak1, Jak2, and Tyk2 phosphorylation of the C VHH IL-10R2 deletion variants in Ba/F3 cells. As shown in Fig. 7E, C VHH IL-10R2(D330) exhibited the previously detected phosphorylation pattern of C VHH IL-10R2 stimulated with 100 ng/ml mCherry-Fc. Interestingly, also C VHH IL-10R2 (D310) showed only some STAT3, Jak1, Jak2, and Tyk2 phosphorylation, whereas the activation of ERK1/2 was lost. Most likely this explains why activation of this receptor fails to induce cellular proliferation. Stimulation of C VHH IL-10R2(D280) and C VHH IL-10R2(D255) with mCherry-Fc did not result in activation of signal transduction due to defective Jak activation (Fig.  7E). These results suggest that the amino acid residues from 310 to 330 but not from 330 to the C terminus are critically needed for signal transduction of homodimeric C VHH IL-10R2. Interestingly, slightly reduced Jak phosphorylation was observed for C VHH IL-10R2(D330) and C VHH IL-10R2(D310), whereas larger deletions within the C VHH IL-10R2 result in Synthetic cytokine receptors complete abrogation of Jak phosphorylation, suggesting that the binding sites of Jak1, Jak2, and Tyk2 are mainly located between amino acid residues 280 and 310.
Amino acid exchanges of prolines 320 and 323 within the intracellular domain of the IL-10R2 only minimally reduce signaling capacity We identified a putative PXXP motif within the intracellular domain of the IL-10R2, which might be involved in noncanonical induction of signal transduction (Fig. 8A) and mutated proline 320 and proline 323 into alanine (C VHH IL-10R2(P320A, P323A), C VHH IL-10R2(P320A,P323A,D330) (Fig. S6A) and generated stably transduced Ba/F3/gp130 cells (Fig. S6B). Proliferation of Ba/F3/gp130/C VHH IL-10R2(P320A,P323A) and Ba/F3/gp130/C VHH IL-10R2(P320A,P323A,D330) cells was, however, still induced by mCherry-Fc, albeit with a slightly higher EC 50 of 73.2 and 71.2 ng/ml, respectively (Fig. 8, B and C), which was also reflected in the slightly reduced induction of the target gene Pim-1 as compared with WT IL-10R2 (Fig.  8D). Also signal transduction was only minimally affected by mutation of the two prolines, and if there was any, then STAT3 phosphorylation was slightly reduced (Fig. 8E). In conclusion, prolines 320 and 323 within the PXXP motif had no effect on signal transduction of homodimeric C VHH IL-10R2.

Discussion
Among the switchable synthetic cytokine receptors, the SyC-yRs belongs to a new class of fully synthetic cytokine systems featuring combinations of synthetic ligands and synthetic receptors (7). In this study, we present three major findings. First, we show that the SyCyR-principle can be adopted to classes of cytokine receptors other than the IL-6/IL-12 cytokine family. In detail, we generated biologically active synthetic cytokine receptors for IL-22. Our data suggest that replacement of the extracellular domain of cytokine receptors by nanobodies as ligand-binding entities is applicable to a wide range of cytokine receptors with associated tyrosine kinases. Therefore, our findings are in good agreement with other studies, using natural cytokines and chimeric cytokine receptors or so-called synthekines, that receptors with associated kinases can be widely combined irrespective of family membership (11,30,31). Second, even though IL-22 signals via a heterodimeric receptor complex consisting of IL-22Ra1 and IL-10R2, we showed that also homodimers of synthetic IL-10R2 but not of IL-22Ra1 are biologically active. Third, SyCyRs for gp130 can form biologically active receptor complexes with synthetic IL-22Ra1 and IL-10R2.
Both receptors for IL-22 belong to the class II cytokine receptor family (32). Long-chain IL-22Ra1 is mainly associated with Jak1, whereas the short-chain IL-10R2 was considered to be mainly associated with Tyk2 (tyrosine kinase 2) (32). The intracellular domain of the long-chain cytokine receptor IL-22Ra1 is 346 amino acid residues long, and apart from predicted box 1 (aa 255-262) and box 2 (aa 282-287) binding sites for Jak1, it also contains a range of canonical activation sites for signal transduction (e.g. STATs and ERK1/2) (25). We expected that homodimers of G VHH IL-22Ra1 or C VHH IL-22Ra1 would also induce signal transduction as we have observed this for other homodimeric long-chain receptor chains, including gp130 (2), IL-12Rb2 (11), and IL-23R (33). Even though homodimers of this synthetic IL-22Ra1 were biologically inactive, this does not necessary mean that an alternative synthetic cytokine receptor composition would also not result in biologically active homodimers. Importantly, the synthetic cytokine receptor for IL-22Ra1 was on the cell surface, and binding of GFP-Fc was verified by flow cytometry. The intracellular domain of the IL-10R2 is, however, only 76 amino acid residues long, and its main function was considered to be the recruitment and activation of Tyk2 (27). So far it has been suggested for the human IL-10R2 that homodimerization can induce phosphorylation of STAT1 (27). However, classical box 1 and box 2 motifs (34) within the intracellular domain of IL-10R2 could not be identified by amino acid sequence comparison. Receptors with short ICDs, including IL-10R2, often bind their ligands with lower affinity, pair with Tyk2 or Jak3, and typically only minimally contribute to STAT recruitment and activation (28,29). Only in combination with long-chain receptors, such as IL-22Ra1, was IL-10R2 considered to contribute to signal transduction. Indeed, IL-22 initially binds to IL-22Ra1, which increases the affinity for IL-10R2 (35)(36)(37)(38). Therefore, it comes as a surprise that homodimers of IL-10R2 induced signal transduction in Ba/F3/ gp130/C VHH IL-10R2 cells, including phosphorylation of STAT3 and ERK1/2. Moreover, this was not only triggered by the phosphorylation of Jak1 and Tyk2, which were also Figure 5. Analysis of IL-10R2 homodimeric signaling using Ba/F3/gp130/C VHH IL-10R2 cells. A, schematic illustration of mCherry-Fc homodimer binding two C VHH IL-10R2 receptors. This image was created with BioRender. B, proliferation of Ba/F3/gp130/C VHH IL-10R2 cells with increasing concentrations of mCherry-Fc from 0.0001 to 1000 ng/ml. Error bars, S.D. ***, p , 0.001. One representative experiment, with three biological replicates, of three is shown. C, proliferation of Ba/F3/gp130 and Ba/F3/gp130/C VHH IL-10R2 cells incubated without cytokine (2), with 10 ng/ml Hyper-IL-6 or 100 ng/ml mCherry-Fc. For the indicated samples, 10 mM P6 inhibitor was added to the respective cytokine. Error bars, S.D. One representative experiment, with four biological replicates, of three is shown. D, STAT3, ERK1/2, Jak1, Jak2, and Tyk2 activation in Ba/F3/gp130/C VHH IL-10R2 cells treated with 10 ng/ml Hyper-IL-6 or 100 ng/ml mCherry, mCherry-Fc for 120 min. Cells treated with the P6 inhibitor were preincubated with 10 mM P6 for 30 min and then also stimulated for 120 min. Equal amounts of protein (50 mg/lane) were analyzed via specific antibodies detecting phospho-STAT3, -ERK1/2, -Jak1, -Jak2, and -Tyk2 and STAT3, ERK1/2, Jak1, Jak2, and Tyk2. Western blotting data show one representative experiment of three. E, HEK293T cells were co-transfected with cDNAs coding for GFP VHH IL-10R2, GFP VHH IL-10R2 and murine Jak1, GFP VHH IL-10R2 and murine Jak2, or GFP VHH IL-10R2 and murine Tyk2. The kinases were immunoprecipitated by specific antibodies, and Western blotting analysis was performed to detect Myc-tagged GFP VHH IL-10R2. IP, coimmunoprecipitation; L, lysate. One representative experiment of two is shown. activated by canonical IL-22 signaling but also by Jak2, which we did not observe in natural and synthetic IL-22 signaling. Interaction of Jak2 and G VHH IL-10R2 was verified by co-immunoprecipitation. Based on our homodimeric receptor analysis using deletion variants of C VHH IL-10R2, our results suggest that the amino acid residues from 310 to 330 are responsible for downstream signaling via STATs and ERK1/2. However, we were not able to identify single amino acids that are responsible for STAT and ERK1/2 activation within this 20-amino-acidresidue-long region. A putative PXXP motif, which might facilitate binding of SH3 domain proteins such as Grb2 (39,40), which is involved in activation of the Ras/Raf MAPK pathway, is not involved in STAT and ERK activation as we have shown by introduction of point mutations to exchange proline into alanine, whereas SH2 domains, found in Grb2 but also in STATs, bind to phosphorylated tyrosine-containing peptides (pYXXQ motif) (41), which are not present in the IL-10R2. Tyrosine-/ SH2-independent STAT activation was reported previously for the granulocyte colony-stimulating factor receptor (G-CSFR; STAT3) (42), IL-22 receptor (IL-22R; STAT3) (43), and interferon-a/b receptor b-chain (IFNAR2; STAT2) (44) and IL-23R (41), and it was speculated that other cytokine receptors may use a similar mode of STAT3 recruitment (44). However, a consensus sequence from these noncanonical STAT activation modi could not be deduced to date. For G-CSFR, STAT3 is not constitutively associated with a tyrosine-mutated G-CSFR (42), and an intermediate molecule might interact with the C-terminal receptor region, which might contain a phosphotyrosinebinding site for the SH2 domain of STAT3 (42). In the case of IL-23R, also STAT3 is not constitutively associated with a short 17-amino-acid-residue-long internal part of the intracellular domain. Dumoutier et al. (43) reported that also the C-terminally located 84 amino acid residues of IL-22R allow constitutive association with STAT3, most likely via the coiled-coil domain of STAT3. Mutation of all cytoplasmic tyrosine residues of the IL-22R only partially affects STAT3 activation, and receptor preassociation with STAT3 might assure a faster STAT3 activation in cells with lower endogenous STAT3 expression. SH2-independent recruitment of STAT3 might serve to avoid negative feedback by proteins such as suppressor of cytokine signaling 3 (SOCS3), which can compete with STAT factors for phosphotyrosines (45).
Albeit a slightly reduced Jak phosphorylation was seen for C VHH IL-10R2(D310) compared with C VHH IL-10R2(D330), only the larger deletion C VHH IL-10R2(D280) resulted in complete abrogation of Jak phosphorylation, suggesting that the binding sites of Jak1, Jak2, and Tyk2 are mainly located between amino acid residues 280 and 310. We have previously defined the noncanonical Jak binding sites in the long-chain IL-23R (46), and others have found that a classical proline-rich box 1 motif does not occur in IFNAR1 (47).
Finally, we show that both IL-22 receptors are able to form biologically active receptor complexes with the IL-6 signaltransducing receptor gp130, which appears to be a common feature of interchangeability among cytokine receptors with associated kinases between different receptor families (2).
Whereas IL-10R2 is widely expressed in cells throughout the body, IL-22Ra1 is expressed predominantly in epithelial tissues. In the intestinal epithelium, IL-22 is responsible for immune homeostasis as well as wound healing (18) and plays an important role in the pathogenesis of many intestinal diseases. Therefore, IL-22 is an interesting therapeutic target for many gastrointestinal diseases (48). Our synthetic IL-22 receptor combination might be a useful tool to study the pro-and anti-inflammatory responses in certain cell types. We are aware that the overexpression of cytokine receptors might cause some artifacts in signaling. Therefore, the next step is to use the endogenous promoter by integration of the synthetic receptors in the endogenous gene loci by CRISPR-Cas9 technology. Moreover, transgenic mice with a Cre-inducible SyCyR for IL-22 may allow the cell type-specific dissection of IL-22 function in tissues and disease states. In general, the modular nature of the synthetic GFP and mCherry ligands allows an exact composition of the receptor stoichiometry, to facilitate tailor-made receptor compositions as shown here for IL-22 and IL-6 crosstalk. In general, this technology might also be used to support CAR T-cell therapies by tailor-made synthetic receptors either supporting or suppressing the activity of CAR T cells.

Construction of SyCyRs and synthetic fluorescent ligands
SyCyR pcDNA3.1 expression plasmids were generated by fusion of coding sequence for the IL-11R signal peptide (Q14626, aa 1-22), a Myc tag (EQKLISEEDL; SyCyRs containing GFP VHH ), or a FLAG tag (DYKDDDDK) and HA tag (YPYDVPDYA; SyCyRs containing mCherry VHH ) followed by a nanobody (GFP VHH or mCherry VHH ), some residues of the extracellular domain (ECD), the complete transmembrane (TMD), and the complete intracellular domain (ICD) of the respective cytokine receptor. For the IL-22Ra-SyCyR, the coding cDNA consists of 10 aa of the ECD and the complete TMD and ICD. The cDNA for the IL-10R2-SyCyR is composed of 10 aa of the ECD and the complete TMD and ICD. The gp130-SyCyR is made up of 13 aa from the ECD and the complete TMD and ICD. Deletion variants of the IL-10R2-SyCyR D330, D310, D280, and D255 were generated by amplification of the respective intracellular part via PCR using Phusion ® high-fidelity DNA polymerase (Thermo Fisher Scientific). Mutation of the SH3 domain of the IL-10R2 was generated by site-directed mutagenesis using Phusion ® high-fidelity DNA polymerase (Thermo Fisher Scientific) followed by DpnI digestion of the methylated template DNA. All SyCyRs were inserted into pMOWS-hygro (49) (mCherry VHH ) or pMOWS-puro (50) (GFP VHH ) vectors for stable transfection of Ba/F3-gp130 cells. All generated plasmids were verified by sequencing.

Generation of synthetic ligands
Synthetic ligands (sequences published previously (2, 5)) were stably expressed in CHO K1 cells using neomycin resistance and single clone selection with 1.125 mg/ml G-418 (Genaxxon Bioscience GmbH). Transfected cells were cultivated with G-418 for 2 weeks, and then single clone selection was carried out with 0.5 cells/well. Single colonies were screened for protein expression. One colony was selected for protein expression in roller bottles (IBS Integra Bioscience, Zizers, Switzerland) with 10% low-IgG fetal calf serum (GIBCO ® , Life Technologies) DMEM for 2 months. The supernatant was collected every 3-4 days, and 1 liter of supernatant was used for purification of Fc-tagged proteins using MabSelect TM HiTrap TM columns (GE Healthcare, Chalfront St Giles, UK). Elution was carried out by pH shift using citrate buffer of pH 5.5 and 3.2. Buffer exchange to PBS was achieved using NAP TM -25 columns (GE Healthcare).

Transfection and selection of cells
Ba/F3-gp130 cells were transduced retrovirally using pMOWS plasmids coding for SyCyRs as described previously (41). As the packaging cell line, Phoenix-Eco cells were used. After transduction, cells were grown as described above and supplemented with puromycin (1.5 mg/ml) and/or hygromycin B (1 mg/ml) (Carl Roth, Karlsruhe, Germany).

Cell viability assay
Ba/F3-gp130 cell lines were washed three times with PBS to remove cytokines from the medium. 5 3 10 4 cells were suspended in DMEM containing 10% fetal calf serum, 60 mg/liter penicillin, and 100 mg/ml streptomycin. Cells were cultured for 3 days in a volume of 100 ml with or without cytokine/synthetic ligands and inhibitors. The CellTiter Blue Viability Assay (Promega, Karlsruhe, Germany) was used to determine the approximate number of viable cells by measuring the fluorescence (excitation 560 nm, emission 590 nm) using the Infinite M200 Pro plate reader (Tecan, Crailsheim, Germany). After adding 20 ml/well of CellTiter Blue reagent (point 0), fluorescence was measured approximately every 20 min for up to 2 h. For each condition of an experiment, 3-4 wells were measured. All values were

Stimulation assay
Ba/F3-gp130 cells were washed four times with PBS to remove cytokines and starved in serum-free DMEM for 4 h. P6 inhibitor was added 30 min prior to stimulation. Cells were stimulated for 1 h (or as indicated if other time points) with 100 ng/ml purified protein, harvested, frozen in liquid nitrogen, and then lysed. Cells were lysed for 1 h with buffer containing 10 mM Tris-HCl, pH 7.8, 150 mM NaCl, 0.5 mM EDTA, 0.5% Nonidet P-40, 1 mM sodium vanadate, 10 mM MgCl 2 and a complete, EDTA-free protease inhibitor mixture tablet (Roche Diagnostics, Mannheim, Germany). Protein concentration was determined by a BCA protein assay (Thermo Fisher Scientific) as described by the manufacturer. Protein expression and activation was analyzed as indicated by immunoblotting of 50 mg of each analysis.

Western blotting
50 mg of protein were loaded per lane and separated by SDS-PAGE under reducing conditions and transferred to a polyvinylidene fluoride membrane (Carl Roth). Blotting of membranes was performed with 5% fat-free dried skimmed milk (Carl Roth) in TBS-T (10 mM Tris-HCl pH 7.6, 150 mM NaCl, 0,5% Tween 20) for 4 h. Primary antibodies were diluted in 5% fatfree milk in TBS-T (STAT3, ERK, g-tubulin) or 5% BSA in TBS-T (pSTAT3, pERK, pJak1, Jak1, pJak2, Jak2, pTyk2, Tyk2, SOCS, HA, Myc) and incubated at 4°C overnight. Membranes were washed with TBS-T and then incubated with the secondary peroxidase-conjugated antibodies in 5% fat-free dried skim milk in TBS-T for at least 1 h. Signal detection was achieved using the ECL Prime Western blotting detection reagent (GE Healthcare, Freiburg, Germany) and the Chemo Cam Imager (INTAS Science Imaging Instruments, Göttingen, Germany). For a second round of detection, the membranes were stripped with 62.5% Tris-HCl, pH 6.8, 2% SDS, 0.1% b-mercaptoethanol at 60°C for 30 min and then blocked again in 5% fat-free dried skimmed milk in TBS-T for at least 3 h before using the next primary antibody.

Immunoprecipitation
Immunoprecipitation was performed with HEK293T cells as described (46).

Cell-surface detection of synthetic cytokine receptors
SyCyR expression of stably transfected Ba/F3-gp130 cells was detected by specific antibodies. Cells were washed in FACS buffer (PBS, 1% BSA) and then resuspended in 50 ml of FACS buffer containing the indicated specific primary antibody (Myc 1:100, HA 1:1000). After incubation of at least 1 h at room temperature, cells were washed and then resuspended in 50 ml of FACS buffer containing secondary antibody Alexa Fluor 488conjugated Fab goat anti-rabbit IgG (catalog no. A11070; 1:500) and incubated for 1 h at room temperature. Cells were washed and resuspended in 500 ml of FACS buffer and analyzed by flow cytometry (BD FACSCanto II flow cytometer, BD Biosciences). Data were evaluated using FlowJo_V10 (FlowJo LLC, Ashland, OR, USA).

Binding of GFP to GFP VHH
The binding of GFP-Fc to the respective GFP VHH was analyzed by flow cytometry using a BD FACSCanto II flow cytometer (BD Biosciences). Cells were incubated without cytokine or with 5 mg/ml GFP-Fc for 1 h at 37°C with 5% CO 2 . Afterward cells were washed three times with PBS, and binding was detected using the FITC-A channel.

Gene expression by real-time PCR
Cells were washed four times with PBS and then starved in serum-free DMEM for 4 h. They were stimulated with 100 ng/ ml for 120 min as indicated, harvested, and frozen in liquid nitrogen. RNA isolation was carried out as described above. RNA concentration was determined by a NanoDrop 2000c spectrophotometer (Thermo Scientific, Waltham, MA, USA) and adjusted to 50 ng/ml for all samples. The expression of specific genes was determined by usage of the iTaq TM Universal SYBR Green One-Step Kit (Bio-Rad) as described previously (51). The expression level of Pim-1 was normalized to the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (Gapdh) for relative quantification and calculated using the D(Ct 1 ) method. The expression level of target genes was determined by AB17500 Real-Time PCR System (Thermo Scientific, Waltham, MA, USA). The following primer pairs were used in this study: GAPDH, fw 59 (GAAGGGCTCATGACCACAGT) and rev 59 (CATTGTCATACCAGGAAATGAGCT); Pim-1, fw 59

Data availability
The data of this study are available within the paper. Gene expression raw data have been deposited in the Gene Expression Omnibus (GEO) with the accession number GSE150919. Figure 11. Analysis of cross-talk between the IL-22Ra1 and gp130 SyCyR. A, schematic illustration of GFP-mCherry heterodimer binding to C VHH IL-22Ra1 (red and brown) and G VHH gp130 (green and purple) SyCyR to induce signal transduction. This image was created with BioRender. B, proliferation of Ba/F3/gp130/G VHH gp130/C VHH IL-22Ra1 cells incubated without cytokine (2), with 10 ng/ml Hyper-IL-6 or 100 ng/ml GFP, mCherry, 2xGFP, 2xmCherry, or GFP-mCherry. Error bars, S.D. ***, p , 0.001. One representative experiment, with three biological replicates, of three is shown. C, proliferation of Ba/F3/gp130/G VHH gp130/C VHH IL-22Ra1 cells incubated with increasing concentrations of GFP-mCherry from 0.1 to 2000 ng/ml. Error bars, S.D. ***, p , 0.001. One representative experiment, with four biological replicates, of three is shown. D, STAT3 activation of Ba/F3/gp130/ G VHH gp130/C VHH IL-22Ra1 cells treated with 10 ng/ml Hyper-IL-6 or 100 ng/ml 2xGFP, 2xmCherry, or GFP-mCherry. One representative experiment of three is shown.