Cytokines modulate the immune response by activating a common JAK/STAT signaling pathway upon cell surface receptor dimerization/oligomerization (Gorby et al., 2018; Stroud and Wells, 2004; Wang et al., 2009; Wells et al., 1993). A conundrum in the field pertains to how biological specificity is achieved in the cytokine system by using such reduced number of signaling intermediaries, that is four JAKs and seven STATs (Murray, 2007; Schindler et al., 2007). Indeed, there are numerous examples in the literature where cytokines activating the same STATs in CD4 T cells, for example IL-6 and IL-10 (Grotzinger et al., 1997; Walter, 2004), produce opposite responses, that is pro-inflammatory vs anti-inflammatory responses respectively (Hunter and Jones, 2015; Wilson et al., 2005).

In recent years, a number of studies in multiple cytokine systems have shown that cytokine signaling is not an ‘all or none’ phenomenon and can be modulated by alterations of cytokine-receptor binding properties (Spangler et al., 2015). Changes in cytokine-receptor binding kinetics and strength were shown to play a crucial role in defining type I and type III interferons biological potencies (Mendoza et al., 2017; Pestka, 2007; Piehler et al., 2012; Subramaniam et al., 1995). A mutation in erythropoietin (Epo) found in humans, which reduces its binding affinity for its receptor (EpoR), was shown to bias signaling output by EpoR and caused severe anemia in human patients (Kim et al., 2017). Biased EpoR signaling was also achieved using surrogate Epo ligands that altered the receptor binding topology (Moraga et al., 2015b). Cross-reactive cytokine-receptor systems, where shared receptors engage multiple cytokines and elicit differential responses is another example where receptor binding properties influence signaling and activity, for example the IL-4/IL-13 system (Heller et al., 2008; LaPorte et al., 2008), IL-2/IL-15 system (Ring et al., 2012; Rochman et al., 2009; Waldmann, 2006) and the IL-6 family system (Wang et al., 2009). Viruses often encode for cytokine-like proteins that bind cytokine receptors with altered binding properties, providing them with means to fine-tune the immune response to their own advantage (Boulanger et al., 2004; Walter, 2004). All these examples strongly argue in favour of cytokine-receptor binding parameters contributing to regulate signaling, however a model providing molecular bases for signaling biased by cytokine receptor is missing.

Biased signaling is not a unique feature of the cytokine family. G-protein-coupled receptors (GPCRs), which contain seven-transmembrane domains, are the classical system where biased signaling was first described (Hilger et al., 2018; Wootten et al., 2018). In this system, different ligands binding a common receptor can trigger differential signaling programs by instructing specific allosteric changes in the transmembrane α-helices of the receptor (Hilger et al., 2018; Wootten et al., 2018). However, this mechanism is more difficult to imagine for cytokine receptors where the transmembrane (TM) region contributes less significantly to signaling. Cytokine receptor chimeras where their TM has been swapped by that of other receptors still trigger signaling (Sharma et al., 2016). How can cytokine receptors trigger biased signaling responses then? A common feature to all cytokine systems is that upon ligand stimulation cytokine-receptor complexes traffic to intracellular compartments, where they are often degraded, contributing to switching off the response (Becker et al., 2010; Bulut et al., 2011; Claudinon et al., 2007; German et al., 2011; Keeler et al., 2007; Shah et al., 2006). However, a complex positive regulatory role of endocytosis in cytokine signaling has emerged (Becker et al., 2010; Cendrowski et al., 2016; Fallon and Lauffenburger, 2000; Marchetti et al., 2006; Sarkar et al., 2002). Several studies recently suggested a novel role for the endosomal compartment in stabilizing cytokine receptor dimers by enhancing local receptor concentrations (Gandhi et al., 2014; Moraga et al., 2015a), thus contributing to signaling fitness even at low complex stabilities. In agreement with this model, mutations on cytokine receptors that alter their intracellular traffic can result in activation of novel or deregulated signaling programs causing disease (Reddy et al., 1996). Furthermore, activated JAK/STAT proteins have been described in endosomes after interferon stimulation, suggesting that signaling continues upon receptor internalization (Payelle-Brogard and Pellegrini, 2010). How changes in cytokine-receptor complex half-life and endosomal trafficking fine-tunes cytokine signaling and biological responses requires further investigation.

Here, using model cell lines and primary human CD4 T cells, we systematically explored how modulation of cytokine-receptor complex stability impacts signaling identity and biological responses, using IL-6 as a model system. IL-6 is a highly pleiotropic cytokine, which critically contributes to mounting the inflammatory response (Grotzinger et al., 1997; Hunter and Jones, 2015; Naka et al., 2002). IL-6 stimulation drives differentiation of Th17 cells (Jones et al., 2010; Kimura and Kishimoto, 2010; Louten et al., 2009), and inhibits the differentiation of Th1 (Diehl and Rincón, 2002) and T regulatory (reg) cells (Kimura and Kishimoto, 2010; Korn et al., 2008). Deregulation of IL-6 levels and activities is often found in human diseases, making IL-6 a very attractive therapeutic target (Hunter and Jones, 2015). IL-6 exerts its immuno-modulatory activities by engaging a hexameric complex comprised of two molecules of IL-6Rα, two molecules of gp130 and two molecules of IL-6, leading to the downstream activation of STAT1 and STAT3 transcription factors (Wang et al., 2009). Using the yeast-surface display engineering platform, we isolated a series of IL-6 variants binding gp130 with different affinities, ranging from wild-type binding affinity to more than 2000-fold enhanced binding. Quantitative signaling and imaging studies revealed that reduction in cytokine-receptor complex stability resulted in differential cytokine-receptor complex dynamics, which ultimately led to activation of biased signaling programs. Low affinity IL-6 variants failed to translocate to intracellular compartments and induce gp130 degradation, triggering STAT3 biased responses. Indeed, inhibition of gp130 intracellular translocation by chemical or genetic blockage of clathrin-mediated trafficking, reduced STAT1 activation levels, without affecting STAT3 activation. Through a series of molecular and cellular assays we demonstrated that STAT1 requires a higher number of phospho-Tyr available in gp130 to reach maximal activation, explaining its enhanced sensitivity to changes in cytokine-receptor dwell-times. The biased signaling programs engaged by the IL-6 variants did not have a linear effect on STAT3 transcriptional activities. Reduced STAT3 activation levels by the low affinity IL-6 variants resulted in graded STAT3 binding to chromatin and gene expression, with some genes exhibiting a high degree of sensitivity to STAT3 activation levels, and other genes being equally induced by all three IL-6 variants. Moreover, IL-6 immuno-modulatory activities exhibited different sensitivity thresholds to changes on STAT activation levels, with Th17 differentiation being induced by all three variants, and inhibition of Treg and Th1 differentiation only robustly promoted by the high affinity variant. Our results provide a molecular model using spatio-temporal dynamics of cytokine-receptor complexes and competitive binding of STATs proteins for phosphorylated tyrosine residues, to explain how cells integrate cytokine signaling signatures into specific biological responses through the establishment of different gene induction thresholds. At the more practical level, our results highlight that manipulation of cytokine-receptor binding parameters via protein engineering is a useful strategy to decouple cytokine functional pleiotropy, a major source of unwanted side effects and toxicity in cytokine-based therapies.

In this study we have engineered IL-6/gp130 binding kinetics to modulate signaling output and decouple IL-6 functional pleiotropy. Two main findings arise from our study: (1) Intracellular traffic dynamics of cytokine-receptor complexes and STATs binding affinities for phospho-Tyr on cytokine receptor ICD act synergistically to define signaling potency and identity, and (2) cells exhibit different gene induction thresholds in response to cytokine partial agonism, which allow them to modulate their responses. The current work, together with previous studies describing signaling tuning in other cytokines systems (Ho et al., 2017; Kim et al., 2017; Moraga et al., 2015b; Mitra et al., 2015), outline a general strategy to design cytokine partial agonists and decouple cytokine functional pleiotropy by modulating cytokine-receptor binding kinetics.

All our IL-6 variants must dimerize gp130 to some extent, since they all trigger signaling. However, our single molecule TIRF data show that low affinity variants struggle to promote detectable gp130 dimerization. These data suggest that non-detectable short-lived IL-6/gp130 complexes can partially engage signaling, but fail to trigger a full response, evoking a kinetic proof reading model. A kinetic proof reading model has been previously proposed for other ligand-receptor systems, including the T cell receptor system (TCR) (McKeithan, 1995) and more recently Receptor Tyrosine Kinases (RTKs) (Zinkle and Mohammadi, 2018). In these two systems, changes on ligand-receptor complex dwell-times induce phosphorylation of different Tyr pools in the receptors ICDs, ultimately recruiting and activating different signaling effectors (Acuto et al., 2008; Lemmon and Schlessinger, 2010). More recently, this model was used to explain biased signaling triggered by an EPO mutant (Kim et al., 2017). However, cytokine receptors differ significantly from RTKs and the TCR. While in these latter receptor systems, activation of different signaling effectors is clearly assigned to phosphorylation of specific Tyr in their ICDs, this is not generally true for cytokine receptors. Often only one or two Tyr in the cytokine receptor ICDs are required for signal activation (Cheng et al., 2011; Schmitz et al., 2000; Zhao et al., 2008).

How can short-lived cytokine-receptor complexes engage different signaling effectors that compete for a single phospho-Tyr? Our study provides new molecular evidences suggesting that STAT activation by cytokine-receptor complexes is governed by a kinetic discrimination mechanism. STATs compete for Tyr in receptors ICDs, thus making them sensitive to changes in cytokine-receptor complex dwell-times. STATs binding with low affinity to phospho-Tyr require longer-lived cytokine-receptor complexes and higher ligand doses to reach maximal activation. In agreement with this model, a previous study reported STAT1 and STAT3 binding with different affinities to phospho-Tyr in gp130 ICD (Wiederkehr-Adam et al., 2003). Moreover, using chimeric receptors and siRNA and overexpression approaches, we show that binding affinity of STAT proteins for phospho-Tyr in receptors ICDs defines signaling amplitude and identity by IL-6. Importantly, this is not the first evidence of STATs competing for phospho-Tyr. IFNα2 activates all STATs molecules, which can be abrogated by a single Tyr mutation in the IFNAR2 ICD, suggesting STAT competition (Zhao et al., 2008). Furthermore, modulation of STAT protein levels has been described to change signaling specificity by cytokines. IFNγ priming, which result in enhanced STAT1 protein levels, shift the IL-10 response from STAT3 activation to STAT1 activation (Herrero et al., 2003). In cells lacking STAT3, IL-6 switches to STAT1 activation, producing IFNγ-like responses (Costa-Pereira et al., 2002).

An alternative model that could explain our observations is that our IL-6 variants exhibit altered gp130 binding topology. Previous studies have shown that changes on cytokine-receptor binding topology, achieved using cytokine surrogate ligands, resulted in biased signaling programs (Livnah et al., 1998; Moraga et al., 2015b). We cannot formally exclude this possibility for our engineered IL-6 variants. The mutations introduced on the IL-6 variants to enhance their affinity for gp130 could have impacted their receptor binding topologies and therefore their signaling properties. However, we believe this possibility is unlikely. First, the mutations engineered in our variants are located exclusively on the IL-6-gp130 site-2 binding interface, leaving the cytokine backbone untouched and therefore greatly reducing the possibility of alterations on ligand-receptor binding topology. Second, there is a large body of literature showing that mutations in cytokine-receptor binding interfaces do not alter receptor binding topology (Ho et al., 2017; Kim et al., 2017; Mendoza et al., 2017; Mitra et al., 2015; Moraga et al., 2015a). Indeed, to our knowledge, there is not a single example in the literature of a cytokine mutant binding with alternative topology to its receptor. Third, our TIRF microscopy experiments confirm a canonical stoichiometry of the complex formed by the different IL-6 variants, excluding the formation of higher order oligomers that could result from alternative binding topologies (Figure 3c, and Figure 3—figure supplement 1d). Overall, our data support a kinetic proof reading model for cytokine signaling, where cytokine-receptor dwell times determine signaling output by kinetically discriminating activation of STAT molecules based on their binding affinity for phospho-Tyr on cytokine receptors ICDs. One key requirement for kinetic proof reading is the existence of an out-of-equilibrium process and rate-limiting kinetic intermediates (Huang et al., 2016; Huang et al., 2019). Importantly, the IL-6 system also follows this general principle. Internalization of the IL-6-gp130 signaling complex acts as a critical event that breaks the equilibrium by irreversibly capturing intact signaling complexes into endosomes. We have recently quantitatively described this effect in a systematic study with engineered IL-13 variants (Moraga et al., 2015a), showing the critical role of interaction rate constants in such a non-equilibrium process. For the engineered gp130 agonists, we clearly observed that endocytosis is altered, which was presumably caused by differential stability of ligand-receptor interaction (i.e. the off-rate), further supporting kinetic proof reading as the main mechanism for cytokine signaling diversification.

Initially thought to contribute to cytokine signaling shutdown, the endosomal compartment has emerged in recent years as a signaling hub, not only in cytokines (Becker et al., 2010; Bulut et al., 2011; Claudinon et al., 2007; German et al., 2011; Keeler et al., 2007; Shah et al., 2006) but also in other ligand-receptor systems (Villaseñor et al., 2016). Previous studies showed that activated JAK and STATs molecules are found in endosomes upon cytokine stimulation (Payelle-Brogard and Pellegrini, 2010), suggesting that cytokine-receptor complexes exhibit a signaling continuum from the cell surface to intracellular compartments. In agreement with this model, recent studies showed that cytokine-receptor complexes traffic to the endosomal compartment, where they are stabilized, contributing to signaling fitness (Gandhi et al., 2014; Moraga et al., 2015a). Only short-lived complexes that fail to traffic to intracellular compartments trigger diminished signaling output. Our data fully support this model and expand it by showing that short-lived complexes that fail to traffic to the endosomal compartment engage a biased signaling program. Low affinity IL-6 variants, which did not induce internalization of gp130, activated more efficiently STAT3 than STAT1. It is possible that the stabilization of short-lived IL-6-gp130 complexes in endosomes provides them with the necessary extra time to activate secondary pathways that engage the receptor with low affinity. An alternative possibility would be that the high density of cytokine receptor phospho-Tyr motifs found in endosomes, as a consequence of the accumulation of cytokine-receptor complexes in the confined endosomal space (Moraga et al., 2015a), could diminish the advantage exhibited by signaling intermediates binding with high affinity cytokine receptors, therefore favouring the activation of secondary low affinity signaling molecules. Further studies will be required to understand kinetics of signal activation by cytokines in intracellular compartments. An open question pertains to whether the intracellular localization of STATs influence signaling by long- or short-lived cytokine receptor complexes. Early studies described the localization of STAT3 in intracellular membranes, while other STATs association with intracellular membranes is not so well described (Shah et al., 2006). Whether signaling by cytokines can be further engineered by modulation of the intracellular localization of STATs requires further investigation.

In the current study, we show that different genes downstream of IL-6 signaling exhibit different thresholds of activation that can be exploited by IL-6 partial agonists to decouple IL-6 immuno-modulatory activities. How these activation thresholds are established is not clear. STAT proteins face two points where the law of mass action influences their responses the most. The first one pertains to the binding of STATs to phospho-Tyr in the receptors ICDs, and as discussed above, contributes to define signaling potency and identity by cytokine-receptor complexes. The second point is found when activated STATs bind specific Gamma interferon Activated Sequences (GAS) motifs in the promoters of responsive genes. GAS sequences, although conserved, exhibit degrees of degeneracy that allow them to bind STATs molecules with different affinities (Bonham et al., 2013; Ehret et al., 2001; Horvath et al., 1995). In addition, different number of GAS sequences are found in different responsive promoters. In principle, the combination of STAT binding affinities for GAS sequences and the number of GAS sequence present in the promoters could generate different gene induction thresholds. In agreement with this model, we identified chromatin regions through our STAT3 Chip-Seq studies, that bound STAT3 with different efficiencies, with regions where STAT3 binding was diminished by changes in STAT3 activation levels, and regions where efficient STAT3 binding was detected in all conditions tested. When we analysed the number of GAS sequences and their motifs under those regions, we could detect that genes that were more sensitive to changes in STAT3 phosphorylation presented higher number of GAS motifs than those more resistant. Overall our data support a kinetic-proof reading model for cytokine signaling, whereby cytokine-receptor dwell time and STAT binding affinities for phospho-Tyr on receptors ICDs define potency and identity of cytokine signaling signatures. As the number of STAT molecules activated by partial agonists increases, additional GAS binding motifs are engaged in promoters with multiple GAS binding sites, triggering the induction of graded gene expression responses. In principle, engineering of cytokine-receptor binding kinetics could rescue cytokine-based therapies, by decreasing cytokine functional pleiotropy and toxicity.

Sequencing data have been deposited in GEO under accession number code GSE130810.

1. 1. Martinez-Fabregas J
   2. Wilmes S
   3. Wang L
   4. Hafer M
   5. Pohler E
   6. Lokau J
   7. Garbers C
   8. Cozzani A
   9. Piehler J
   10. Kazemian M
   11. Mitra S
   12. Moraga I

   (2019) NCBI Gene Expression Omnibus

   ID GSE130810. Kinetics of cytokine receptor trafficking determine signaling and functional selectivity.

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE130810

1. 1. Acuto O
   2. Di Bartolo V
   3. Michel F

   (2008) Tailoring T-cell receptor signals by proximal negative feedback mechanisms

   Nature Reviews Immunology 8:699–712.

   https://doi.org/10.1038/nri2397

   • PubMed
   • Google Scholar
2. 1. Becker V
   2. Schilling M
   3. Bachmann J
   4. Baumann U
   5. Raue A
   6. Maiwald T
   7. Timmer J
   8. Klingmüller U

   (2010) Covering a broad dynamic range: information processing at the erythropoietin receptor

   Science 328:1404–1408.

   https://doi.org/10.1126/science.1184913

   • PubMed
   • Google Scholar
3. 1. Boder ET
   2. Wittrup KD

   (1997) Yeast surface display for screening combinatorial polypeptide libraries

   Nature Biotechnology 15:553–557.

   https://doi.org/10.1038/nbt0697-553

   • PubMed
   • Google Scholar
4. 1. Bonham AJ
   2. Wenta N
   3. Osslund LM
   4. Prussin AJ
   5. Vinkemeier U
   6. Reich NO

   (2013) STAT1:dna sequence-dependent binding modulation by phosphorylation, protein:protein interactions and small-molecule inhibition

   Nucleic Acids Research 41:754–763.

   https://doi.org/10.1093/nar/gks1085

   • PubMed
   • Google Scholar
5. 1. Boulanger MJ
   2. Chow DC
   3. Brevnova E
   4. Martick M
   5. Sandford G
   6. Nicholas J
   7. Garcia KC

   (2004) Molecular mechanisms for viral mimicry of a human cytokine: activation of gp130 by HHV-8 interleukin-6

   Journal of Molecular Biology 335:641–654.

   https://doi.org/10.1016/j.jmb.2003.10.070

   • PubMed
   • Google Scholar
6. 1. Bulut GB
   2. Sulahian R
   3. Ma Y
   4. Chi NW
   5. Huang LJ

   (2011) Ubiquitination regulates the internalization, endolysosomal sorting, and signaling of the erythropoietin receptor

   Journal of Biological Chemistry 286:6449–6457.

   https://doi.org/10.1074/jbc.M110.186890

   • PubMed
   • Google Scholar
7. 1. Cendrowski J
   2. Mamińska A
   3. Miaczynska M

   (2016) Endocytic regulation of cytokine receptor signaling

   Cytokine & Growth Factor Reviews 32:63–73.

   https://doi.org/10.1016/j.cytogfr.2016.07.002

   • PubMed
   • Google Scholar
8. 1. Cheng G
   2. Yu A
   3. Malek TR

   (2011) T-cell tolerance and the multi-functional role of IL-2R signaling in T-regulatory cells

   Immunological Reviews 241:63–76.

   https://doi.org/10.1111/j.1600-065X.2011.01004.x

   • PubMed
   • Google Scholar
9. 1. Claudinon J
   2. Monier MN
   3. Lamaze C

   (2007) Interfering with interferon receptor sorting and trafficking: impact on signaling

   Biochimie 89:735–743.

   https://doi.org/10.1016/j.biochi.2007.03.014

   • PubMed
   • Google Scholar
10. 1. Costa-Pereira AP
    2. Tininini S
    3. Strobl B
    4. Alonzi T
    5. Schlaak JF
    6. Is'harc H
    7. Gesualdo I
    8. Newman SJ
    9. Kerr IM
    10. Poli V

    (2002) Mutational switch of an IL-6 response to an interferon-gamma-like response

    PNAS 99:8043–8047.

    https://doi.org/10.1073/pnas.122236099

    • PubMed
    • Google Scholar
11. 1. Diehl S
    2. Rincón M

    (2002) The two faces of IL-6 on Th1/Th2 differentiation

    Molecular Immunology 39:531–536.

    https://doi.org/10.1016/S0161-5890(02)00210-9

    • PubMed
    • Google Scholar
12. 1. Ehret GB
    2. Reichenbach P
    3. Schindler U
    4. Horvath CM
    5. Fritz S
    6. Nabholz M
    7. Bucher P

    (2001) DNA binding specificity of different STAT proteins. Comparison of in vitro specificity with natural target sites

    The Journal of Biological Chemistry 276:6675–6688.

    https://doi.org/10.1074/jbc.M001748200

    • PubMed
    • Google Scholar
13. 1. Fallon EM
    2. Lauffenburger DA

    (2000) Computational model for effects of ligand/receptor binding properties on interleukin-2 trafficking dynamics and T cell proliferation response

    Biotechnology Progress 16:905–916.

    https://doi.org/10.1021/bp000097t

    • PubMed
    • Google Scholar
14. 1. Fischer M
    2. Goldschmitt J
    3. Peschel C
    4. Brakenhoff JP
    5. Kallen KJ
    6. Wollmer A
    7. Grötzinger J
    8. Rose-John S

    (1997) I. A bioactive designer cytokine for human hematopoietic progenitor cell expansion

    Nature Biotechnology 15:142–145.

    https://doi.org/10.1038/nbt0297-142

    • PubMed
    • Google Scholar
15. 1. Gandhi H
    2. Worch R
    3. Kurgonaite K
    4. Hintersteiner M
    5. Schwille P
    6. Bökel C
    7. Weidemann T

    (2014) Dynamics and interaction of interleukin-4 receptor subunits in living cells

    Biophysical Journal 107:2515–2527.

    https://doi.org/10.1016/j.bpj.2014.07.077

    • PubMed
    • Google Scholar
16. 1. Garbers C
    2. Thaiss W
    3. Jones GW
    4. Waetzig GH
    5. Lorenzen I
    6. Guilhot F
    7. Lissilaa R
    8. Ferlin WG
    9. Grötzinger J
    10. Jones SA
    11. Rose-John S
    12. Scheller J

    (2011) Inhibition of classic signaling is a novel function of soluble glycoprotein 130 (sgp130), which is controlled by the ratio of interleukin 6 and soluble interleukin 6 receptor

    Journal of Biological Chemistry 286:42959–42970.

    https://doi.org/10.1074/jbc.M111.295758

    • PubMed
    • Google Scholar
17. 1. Gearing DP
    2. Ziegler SF
    3. Comeau MR
    4. Friend D
    5. Thoma B
    6. Cosman D
    7. Park L
    8. Mosley B

    (1994) Proliferative responses and binding properties of hematopoietic cells transfected with low-affinity receptors for leukemia inhibitory factor, oncostatin M, and ciliary neurotrophic factor

    PNAS 91:1119–1123.

    https://doi.org/10.1073/pnas.91.3.1119

    • PubMed
    • Google Scholar
18. 1. Gerhartz C
    2. Heesel B
    3. Sasse J
    4. Hemmann U
    5. Landgraf C
    6. Schneider-Mergener J
    7. Horn F
    8. Heinrich PC
    9. Graeve L

    (1996) Differential activation of acute phase response factor/STAT3 and STAT1 via the cytoplasmic domain of the interleukin 6 signal transducer gp130. I. definition of a novel phosphotyrosine motif mediating STAT1 activation

    The Journal of Biological Chemistry 271:12991–12998.

    https://doi.org/10.1074/jbc.271.22.12991

    • PubMed
    • Google Scholar
19. 1. German CL
    2. Sauer BM
    3. Howe CL

    (2011) The STAT3 beacon: il-6 recurrently activates STAT 3 from endosomal structures

    Experimental Cell Research 317:1955–1969.

    https://doi.org/10.1016/j.yexcr.2011.05.009

    • Google Scholar
20. 1. Gil MP
    2. Bohn E
    3. O'Guin AK
    4. Ramana CV
    5. Levine B
    6. Stark GR
    7. Virgin HW
    8. Schreiber RD

    (2001) Biologic consequences of Stat1-independent IFN signaling

    PNAS 98:6680–6685.

    https://doi.org/10.1073/pnas.111163898

    • PubMed
    • Google Scholar
21. 1. Gonnord P
    2. Blouin CM
    3. Lamaze C

    (2012) Membrane trafficking and signaling: two sides of the same coin

    Seminars in Cell & Developmental Biology 23:154–164.

    https://doi.org/10.1016/j.semcdb.2011.11.002

    • PubMed
    • Google Scholar
22. 1. Gorby C
    2. Martinez-Fabregas J
    3. Wilmes S
    4. Moraga I

    (2018) Mapping determinants of cytokine signaling via protein engineering

    Frontiers in Immunology 9:2143.

    https://doi.org/10.3389/fimmu.2018.02143

    • PubMed
    • Google Scholar
23. 1. Grotzinger J
    2. Kurapkat G
    3. Wollmer A
    4. Kalai M
    5. Rose-John S

    (1997) The family of the IL-6-Type cytokines: specificity and promiscuity of the receptor complexes

    Proteins: Structure, Function, and Genetics 27:96–109.

    https://doi.org/10.1002/(SICI)1097-0134(199701)27:1<96::AID-PROT10>3.0.CO;2-D

    • Google Scholar
24. 1. Hamers-Casterman C
    2. Atarhouch T
    3. Muyldermans S
    4. Robinson G
    5. Hammers C
    6. Songa EB
    7. Bendahman N
    8. Hammers R

    (1993) Naturally occurring antibodies devoid of light chains

    Nature 363:446–448.

    https://doi.org/10.1038/363446a0

    • Google Scholar
25. 1. Heink S
    2. Yogev N
    3. Garbers C
    4. Herwerth M
    5. Aly L
    6. Gasperi C
    7. Husterer V
    8. Croxford AL
    9. Möller-Hackbarth K
    10. Bartsch HS
    11. Sotlar K
    12. Krebs S
    13. Regen T
    14. Blum H
    15. Hemmer B
    16. Misgeld T
    17. Wunderlich TF
    18. Hidalgo J
    19. Oukka M
    20. Rose-John S
    21. Schmidt-Supprian M
    22. Waisman A
    23. Korn T

    (2017) Trans-presentation of IL-6 by dendritic cells is required for the priming of pathogenic T_H17 cells

    Nature Immunology 18:74–85.

    https://doi.org/10.1038/ni.3632

    • PubMed
    • Google Scholar
26. 1. Heller NM
    2. Qi X
    3. Junttila IS
    4. Shirey KA
    5. Vogel SN
    6. Paul WE
    7. Keegan AD

    (2008) Type I IL-4Rs selectively activate IRS-2 to induce target gene expression in macrophages

    Science Signaling 1:ra17.

    https://doi.org/10.1126/scisignal.1164795

    • PubMed
    • Google Scholar
27. 1. Herrero C
    2. Hu X
    3. Li WP
    4. Samuels S
    5. Sharif MN
    6. Kotenko S
    7. Ivashkiv LB

    (2003) Reprogramming of IL-10 activity and signaling by IFN-gamma

    The Journal of Immunology 171:5034–5041.

    https://doi.org/10.4049/jimmunol.171.10.5034

    • PubMed
    • Google Scholar
28. 1. Hilger D
    2. Masureel M
    3. Kobilka BK

    (2018) Structure and dynamics of GPCR signaling complexes

    Nature Structural & Molecular Biology 25:4–12.

    https://doi.org/10.1038/s41594-017-0011-7

    • PubMed
    • Google Scholar
29. 1. Ho CCM
    2. Chhabra A
    3. Starkl P
    4. Schnorr P-J
    5. Wilmes S
    6. Moraga I
    7. Kwon H-S
    8. Gaudenzio N
    9. Sibilano R
    10. Wehrman TS
    11. Gakovic M
    12. Sockolosky JT
    13. Tiffany MR
    14. Ring AM
    15. Piehler J
    16. Weissman IL
    17. Galli SJ
    18. Shizuru JA
    19. Garcia KC

    (2017) Decoupling the functional pleiotropy of stem cell factor by tuning c-Kit signaling

    Cell 168:1018–1052.

    https://doi.org/10.1016/j.cell.2017.02.011

    • Google Scholar
30. 1. Horvath CM
    2. Wen Z
    3. Darnell JE

    (1995) A STAT protein domain that determines DNA sequence recognition suggests a novel DNA-binding domain

    Genes & Development 9:984–994.

    https://doi.org/10.1101/gad.9.8.984

    • PubMed
    • Google Scholar
31. 1. Huang WY
    2. Yan Q
    3. Lin WC
    4. Chung JK
    5. Hansen SD
    6. Christensen SM
    7. Tu HL
    8. Kuriyan J
    9. Groves JT

    (2016) Phosphotyrosine-mediated LAT assembly on membranes drives kinetic bifurcation in recruitment dynamics of the ras activator SOS

    PNAS 113:8218–8223.

    https://doi.org/10.1073/pnas.1602602113

    • PubMed
    • Google Scholar
32. 1. Huang WYC
    2. Alvarez S
    3. Kondo Y
    4. Lee YK
    5. Chung JK
    6. Lam HYM
    7. Biswas KH
    8. Kuriyan J
    9. Groves JT

    (2019) A molecular assembly phase transition and kinetic proofreading modulate ras activation by SOS

    Science 363:1098–1103.

    https://doi.org/10.1126/science.aau5721

    • PubMed
    • Google Scholar
33. 1. Hunter CA
    2. Jones SA

    (2015) IL-6 as a keystone cytokine in health and disease

    Nature Immunology 16:448–457.

    https://doi.org/10.1038/ni.3153

    • Google Scholar
34. 1. Jones GW
    2. McLoughlin RM
    3. Hammond VJ
    4. Parker CR
    5. Williams JD
    6. Malhotra R
    7. Scheller J
    8. Williams AS
    9. Rose-John S
    10. Topley N
    11. Jones SA

    (2010) Loss of CD4⁺ T cell IL-6R expression during inflammation underlines a role for IL-6 trans signaling in the local maintenance of Th17 cells

    Journal of Immunology 184:2130–2139.

    https://doi.org/10.4049/jimmunol.0901528

    • PubMed
    • Google Scholar
35. 1. Keeler C
    2. Jablonski EM
    3. Albert YB
    4. Taylor BD
    5. Myszka DG
    6. Clevenger CV
    7. Hodsdon ME

    (2007) The kinetics of binding human prolactin, but not growth hormone, to the prolactin receptor vary over a physiologic pH range

    Biochemistry 46:2398–2410.

    https://doi.org/10.1021/bi061958v

    • PubMed
    • Google Scholar
36. 1. Kim AR
    2. Ulirsch JC
    3. Wilmes S
    4. Unal E
    5. Moraga I
    6. Karakukcu M
    7. Yuan D
    8. Kazerounian S
    9. Abdulhay NJ
    10. King DS
    11. Gupta N
    12. Gabriel SB
    13. Lander ES
    14. Patiroglu T
    15. Ozcan A
    16. Ozdemir MA
    17. Garcia KC
    18. Piehler J
    19. Gazda HT
    20. Klein DE
    21. Sankaran VG

    (2017) Functional selectivity in cytokine signaling revealed through a pathogenic EPO mutation

    Cell 168:1053–1064.

    https://doi.org/10.1016/j.cell.2017.02.026

    • PubMed
    • Google Scholar
37. 1. Kimura A
    2. Kishimoto T

    (2010) IL-6: regulator of treg/Th17 balance

    European Journal of Immunology 40:1830–1835.

    https://doi.org/10.1002/eji.201040391

    • Google Scholar
38. 1. Kirchhofer A
    2. Helma J
    3. Schmidthals K
    4. Frauer C
    5. Cui S
    6. Karcher A
    7. Pellis M
    8. Muyldermans S
    9. Casas-Delucchi CS
    10. Cardoso MC
    11. Leonhardt H
    12. Hopfner KP
    13. Rothbauer U

    (2010) Modulation of protein properties in living cells using nanobodies

    Nature Structural & Molecular Biology 17:133–138.

    https://doi.org/10.1038/nsmb.1727

    • PubMed
    • Google Scholar
39. 1. Korn T
    2. Mitsdoerffer M
    3. Croxford AL
    4. Awasthi A
    5. Dardalhon VA
    6. Galileos G
    7. Vollmar P
    8. Stritesky GL
    9. Kaplan MH
    10. Waisman A
    11. Kuchroo VK
    12. Oukka M

    (2008) IL-6 controls Th17 immunity in vivo by inhibiting the conversion of conventional T cells into Foxp3+ regulatory T cells

    PNAS 105:18460–18465.

    https://doi.org/10.1073/pnas.0809850105

    • PubMed
    • Google Scholar
40. 1. Krutzik PO
    2. Nolan GP

    (2006) Fluorescent cell barcoding in flow cytometry allows high-throughput drug screening and signaling profiling

    Nature Methods 3:361–368.

    https://doi.org/10.1038/nmeth872

    • PubMed
    • Google Scholar
41. 1. LaPorte SL
    2. Juo ZS
    3. Vaclavikova J
    4. Colf LA
    5. Qi X
    6. Heller NM
    7. Keegan AD
    8. Garcia KC

    (2008) Molecular and structural basis of cytokine receptor pleiotropy in the interleukin-4/13 system

    Cell 132:259–272.

    https://doi.org/10.1016/j.cell.2007.12.030

    • PubMed
    • Google Scholar
42. 1. Lemmon MA
    2. Schlessinger J

    (2010) Cell signaling by receptor tyrosine kinases

    Cell 141:1117–1134.

    https://doi.org/10.1016/j.cell.2010.06.011

    • PubMed
    • Google Scholar
43. 1. Liao W
    2. Schones DE
    3. Oh J
    4. Cui Y
    5. Cui K
    6. Roh TY
    7. Zhao K
    8. Leonard WJ

    (2008) Priming for T helper type 2 differentiation by interleukin 2-mediated induction of interleukin 4 receptor alpha-chain expression

    Nature Immunology 9:1288–1296.

    https://doi.org/10.1038/ni.1656

    • PubMed
    • Google Scholar
44. 1. Livnah O
    2. Johnson DL
    3. Stura EA
    4. Farrell FX
    5. Barbone FP
    6. You Y
    7. Liu KD
    8. Goldsmith MA
    9. He W
    10. Krause CD
    11. Pestka S
    12. Jolliffe LK
    13. Wilson IA

    (1998) An antagonist peptide-EPO receptor complex suggests that receptor dimerization is not sufficient for activation

    Nature Structural Biology 5:993–1004.

    https://doi.org/10.1038/2965

    • PubMed
    • Google Scholar
45. 1. Louten J
    2. Boniface K
    3. de Waal Malefyt R

    (2009) Development and function of TH17 cells in health and disease

    Journal of Allergy and Clinical Immunology 123:1004–1011.

    https://doi.org/10.1016/j.jaci.2009.04.003

    • PubMed
    • Google Scholar
46. 1. Marchetti M
    2. Monier MN
    3. Fradagrada A
    4. Mitchell K
    5. Baychelier F
    6. Eid P
    7. Johannes L
    8. Lamaze C

    (2006) Stat-mediated signaling induced by type I and type II interferons (IFNs) is differentially controlled through lipid microdomain association and clathrin-dependent endocytosis of IFN receptors

    Molecular Biology of the Cell 17:2896–2909.

    https://doi.org/10.1091/mbc.e06-01-0076

    • PubMed
    • Google Scholar
47. 1. McKeithan TW

    (1995) Kinetic proofreading in T-cell receptor signal transduction

    PNAS 92:5042–5046.

    https://doi.org/10.1073/pnas.92.11.5042

    • PubMed
    • Google Scholar
48. 1. Mendoza JL
    2. Schneider WM
    3. Hoffmann H-H
    4. Vercauteren K
    5. Jude KM
    6. Xiong A
    7. Moraga I
    8. Horton TM
    9. Glenn JS
    10. de Jong YP
    11. Rice CM
    12. Garcia KC

    (2017) The IFN-λ-IFN-λR1-IL-10Rβ complex reveals structural features underlying type III IFN functional plasticity

    Immunity 46:379–392.

    https://doi.org/10.1016/j.immuni.2017.02.017

    • Google Scholar
49. 1. Mitra S
    2. Ring AM
    3. Amarnath S
    4. Spangler JB
    5. Li P
    6. Ju W
    7. Fischer S
    8. Oh J
    9. Spolski R
    10. Weiskopf K
    11. Kohrt H
    12. Foley JE
    13. Rajagopalan S
    14. Long EO
    15. Fowler DH
    16. Waldmann TA
    17. Garcia KC
    18. Leonard WJ

    (2015) Interleukin-2 activity can be fine tuned with engineered receptor signaling clamps

    Immunity 42:826–838.

    https://doi.org/10.1016/j.immuni.2015.04.018

    • PubMed
    • Google Scholar
50. 1. Moraga I
    2. Richter D
    3. Wilmes S
    4. Winkelmann H
    5. Jude K
    6. Thomas C
    7. Suhoski MM
    8. Engleman EG
    9. Piehler J
    10. Garcia KC

    (2015a) Instructive roles for cytokine-receptor binding parameters in determining signaling and functional potency

    Science Signaling 8:ra114.

    https://doi.org/10.1126/scisignal.aab2677

    • PubMed
    • Google Scholar
51. 1. Moraga I
    2. Wernig G
    3. Wilmes S
    4. Gryshkova V
    5. Richter CP
    6. Hong WJ
    7. Sinha R
    8. Guo F
    9. Fabionar H
    10. Wehrman TS
    11. Krutzik P
    12. Demharter S
    13. Plo I
    14. Weissman IL
    15. Minary P
    16. Majeti R
    17. Constantinescu SN
    18. Piehler J
    19. Garcia KC

    (2015b) Tuning cytokine receptor signaling by re-orienting dimer geometry with surrogate ligands

    Cell 160:1196–1208.

    https://doi.org/10.1016/j.cell.2015.02.011

    • PubMed
    • Google Scholar
52. 1. Murray PJ

    (2007) The JAK-STAT signaling pathway: input and output integration

    The Journal of Immunology 178:2623–2629.

    https://doi.org/10.4049/jimmunol.178.5.2623

    • PubMed
    • Google Scholar
53. 1. Naka T
    2. Nishimoto N
    3. Kishimoto T

    (2002) The paradigm of IL-6: from basic science to medicine

    Arthritis Research 4:233–242.

    https://doi.org/10.1186/ar565

    • PubMed
    • Google Scholar
54. 1. Payelle-Brogard B
    2. Pellegrini S

    (2010) Biochemical monitoring of the early endocytic traffic of the type I interferon receptor

    Journal of Interferon & Cytokine Research 30:89–98.

    https://doi.org/10.1089/jir.2009.0044

    • PubMed
    • Google Scholar
55. 1. Pestka S

    (2007) The interferons: 50 years after their discovery, there is much more to learn

    Journal of Biological Chemistry 282:20047–20051.

    https://doi.org/10.1074/jbc.R700004200

    • PubMed
    • Google Scholar
56. 1. Piehler J
    2. Thomas C
    3. Garcia KC
    4. Schreiber G

    (2012) Structural and dynamic determinants of type I interferon receptor assembly and their functional interpretation

    Immunological Reviews 250:317–334.

    https://doi.org/10.1111/imr.12001

    • PubMed
    • Google Scholar
57. 1. Reddy CC
    2. Niyogi SK
    3. Wells A
    4. Wiley HS
    5. Lauffenburger DA

    (1996) Engineering epidermal growth factor for enhanced mitogenic potency

    Nature Biotechnology 14:1696–1699.

    https://doi.org/10.1038/nbt1296-1696

    • PubMed
    • Google Scholar
58. 1. Ring AM
    2. Lin JX
    3. Feng D
    4. Mitra S
    5. Rickert M
    6. Bowman GR
    7. Pande VS
    8. Li P
    9. Moraga I
    10. Spolski R
    11. Ozkan E
    12. Leonard WJ
    13. Garcia KC

    (2012) Mechanistic and structural insight into the functional dichotomy between IL-2 and IL-15

    Nature Immunology 13:1187–1195.

    https://doi.org/10.1038/ni.2449

    • PubMed
    • Google Scholar
59. 1. Rochman Y
    2. Spolski R
    3. Leonard WJ

    (2009) New insights into the regulation of T cells by gamma(c) family cytokines

    Nature Reviews Immunology 9:480–490.

    https://doi.org/10.1038/nri2580

    • PubMed
    • Google Scholar
60. 1. Röder F
    2. Lubk A
    3. Wolf D
    4. Niermann T

    (2014) Noise estimation for off-axis electron holography

    Ultramicroscopy 144:32–42.

    https://doi.org/10.1016/j.ultramic.2014.04.002

    • PubMed
    • Google Scholar
61. 1. Sarkar CA
    2. Lowenhaupt K
    3. Horan T
    4. Boone TC
    5. Tidor B
    6. Lauffenburger DA

    (2002) Rational cytokine design for increased lifetime and enhanced potency using pH-activated "histidine switching"

    Nature Biotechnology 20:908–913.

    https://doi.org/10.1038/nbt725

    • PubMed
    • Google Scholar
62. 1. Schindler C
    2. Levy DE
    3. Decker T

    (2007) JAK-STAT signaling: from interferons to cytokines

    Journal of Biological Chemistry 282:20059–20063.

    https://doi.org/10.1074/jbc.R700016200

    • PubMed
    • Google Scholar
63. 1. Schmitz J
    2. Dahmen H
    3. Grimm C
    4. Gendo C
    5. Müller-Newen G
    6. Heinrich PC
    7. Schaper F

    (2000) The cytoplasmic tyrosine motifs in full-length glycoprotein 130 have different roles in IL-6 signal transduction

    The Journal of Immunology 164:848–854.

    https://doi.org/10.4049/jimmunol.164.2.848

    • PubMed
    • Google Scholar
64. 1. Sergé A
    2. Bertaux N
    3. Rigneault H
    4. Marguet D

    (2008) Dynamic multiple-target tracing to probe spatiotemporal cartography of cell membranes

    Nature Methods 5:687–694.

    https://doi.org/10.1038/nmeth.1233

    • PubMed
    • Google Scholar
65. 1. Shah M
    2. Patel K
    3. Mukhopadhyay S
    4. Xu F
    5. Guo G
    6. Sehgal PB

    (2006) Membrane-associated STAT3 and PY-STAT3 in the cytoplasm

    Journal of Biological Chemistry 281:7302–7308.

    https://doi.org/10.1074/jbc.M508527200

    • PubMed
    • Google Scholar
66. 1. Sharma N
    2. Longjam G
    3. Schreiber G

    (2016) Type I interferon signaling is decoupled from specific receptor orientation through lenient requirements of the transmembrane domain

    Journal of Biological Chemistry 291:3371–3384.

    https://doi.org/10.1074/jbc.M115.686071

    • PubMed
    • Google Scholar
67. 1. Spangler JB
    2. Moraga I
    3. Mendoza JL
    4. Garcia KC

    (2015) Insights into cytokine-receptor interactions from cytokine engineering

    Annual Review of Immunology 33:139–167.

    https://doi.org/10.1146/annurev-immunol-032713-120211

    • PubMed
    • Google Scholar
68. 1. Stroud RM
    2. Wells JA

    (2004) Mechanistic diversity of cytokine receptor signaling across cell membranes

    Science Signaling 2004:re7.

    https://doi.org/10.1126/stke.2312004re7

    • Google Scholar
69. 1. Stumhofer JS
    2. Tait ED
    3. Quinn WJ
    4. Hosken N
    5. Spudy B
    6. Goenka R
    7. Fielding CA
    8. O'Hara AC
    9. Chen Y
    10. Jones ML
    11. Saris CJ
    12. Rose-John S
    13. Cua DJ
    14. Jones SA
    15. Elloso MM
    16. Grötzinger J
    17. Cancro MP
    18. Levin SD
    19. Hunter CA

    (2010) A role for IL-27p28 as an antagonist of gp130-mediated signaling

    Nature Immunology 11:1119–1126.

    https://doi.org/10.1038/ni.1957

    • PubMed
    • Google Scholar
70. 1. Subramaniam PS
    2. Khan SA
    3. Pontzer CH
    4. Johnson HM

    (1995) Differential recognition of the type I interferon receptor by interferons tau and alpha is responsible for their disparate cytotoxicities

    PNAS 92:12270–12274.

    https://doi.org/10.1073/pnas.92.26.12270

    • PubMed
    • Google Scholar
71. 1. Tanaka Y
    2. Tanaka N
    3. Saeki Y
    4. Tanaka K
    5. Murakami M
    6. Hirano T
    7. Ishii N
    8. Sugamura K

    (2008) c-Cbl-dependent monoubiquitination and lysosomal degradation of gp130

    Molecular and Cellular Biology 28:4805–4818.

    https://doi.org/10.1128/MCB.01784-07

    • PubMed
    • Google Scholar
72. 1. Villaseñor R
    2. Kalaidzidis Y
    3. Zerial M

    (2016) Signal processing by the endosomal system

    Current Opinion in Cell Biology 39:53–60.

    https://doi.org/10.1016/j.ceb.2016.02.002

    • PubMed
    • Google Scholar
73. 1. Vogelsang J
    2. Kasper R
    3. Steinhauer C
    4. Person B
    5. Heilemann M
    6. Sauer M
    7. Tinnefeld P

    (2008) A reducing and oxidizing system minimizes photobleaching and blinking of fluorescent dyes

    Angewandte Chemie International Edition 47:5465–5469.

    https://doi.org/10.1002/anie.200801518

    • PubMed
    • Google Scholar
74. 1. Waichman S
    2. Bhagawati M
    3. Podoplelova Y
    4. Reichel A
    5. Brunk A
    6. Paterok D
    7. Piehler J

    (2010) Functional immobilization and patterning of proteins by an enzymatic transfer reaction

    Analytical Chemistry 82:1478–1485.

    https://doi.org/10.1021/ac902608a

    • PubMed
    • Google Scholar
75. 1. Waldmann TA

    (2006) The biology of interleukin-2 and interleukin-15: implications for Cancer therapy and vaccine design

    Nature Reviews Immunology 6:595–601.

    https://doi.org/10.1038/nri1901

    • PubMed
    • Google Scholar
76. 1. Walter MR

    (2004) Structural analysis of IL-10 and type I interferon family members and their complexes with receptor

    Advances in Protein Chemistry 68:171–223.

    https://doi.org/10.1016/S0065-3233(04)68006-5

    • PubMed
    • Google Scholar
77. 1. Wang X
    2. Lupardus P
    3. Laporte SL
    4. Garcia KC

    (2009) Structural biology of shared cytokine receptors

    Annual Review of Immunology 27:29–60.

    https://doi.org/10.1146/annurev.immunol.24.021605.090616

    • PubMed
    • Google Scholar
78. 1. Wells JA
    2. Cunningham BC
    3. Fuh G
    4. Lowman HB
    5. Bass SH
    6. Mulkerrin MG
    7. Ultsch M
    8. deVos AM

    (1993) The molecular basis for growth hormone-receptor interactions

    Recent Progress in Hormone Research 48:253–275.

    https://doi.org/10.1016/B978-0-12-571148-7.50013-0

    • PubMed
    • Google Scholar
79. 1. Wiederkehr-Adam M
    2. Ernst P
    3. Müller K
    4. Bieck E
    5. Gombert FO
    6. Ottl J
    7. Graff P
    8. Grossmüller F
    9. Heim MH

    (2003) Characterization of phosphopeptide motifs specific for the src homology 2 domains of signal transducer and activator of transcription 1 (STAT1) and STAT3

    Journal of Biological Chemistry 278:16117–16128.

    https://doi.org/10.1074/jbc.M300261200

    • PubMed
    • Google Scholar
80. 1. Wilmes S
    2. Beutel O
    3. Li Z
    4. Francois-Newton V
    5. Richter CP
    6. Janning D
    7. Kroll C
    8. Hanhart P
    9. Hötte K
    10. You C
    11. Uzé G
    12. Pellegrini S
    13. Piehler J

    (2015) Receptor dimerization dynamics as a regulatory valve for plasticity of type I interferon signaling

    The Journal of Cell Biology 209:579–593.

    https://doi.org/10.1083/jcb.201412049

    • PubMed
    • Google Scholar
81. 1. Wilson EH
    2. Wille-Reece U
    3. Dzierszinski F
    4. Hunter CA

    (2005) A critical role for IL-10 in limiting inflammation during toxoplasmic encephalitis

    Journal of Neuroimmunology 165:63–74.

    https://doi.org/10.1016/j.jneuroim.2005.04.018

    • PubMed
    • Google Scholar
82. 1. Wootten D
    2. Christopoulos A
    3. Marti-Solano M
    4. Babu MM
    5. Sexton PM

    (2018) Mechanisms of signalling and biased agonism in G protein-coupled receptors

    Nature Reviews Molecular Cell Biology 19:638–653.

    https://doi.org/10.1038/s41580-018-0049-3

    • PubMed
    • Google Scholar
83. 1. Wung BS
    2. Ni CW
    3. Wang DL

    (2005) ICAM-1 induction by TNFalpha and IL-6 is mediated by distinct pathways via rac in endothelial cells

    Journal of Biomedical Science 12:91–101.

    https://doi.org/10.1007/s11373-004-8170-z

    • PubMed
    • Google Scholar
84. 1. You C
    2. Marquez-Lago TT
    3. Richter CP
    4. Wilmes S
    5. Moraga I
    6. Garcia KC
    7. Leier A
    8. Piehler J

    (2016) Receptor dimer stabilization by hierarchical plasma membrane microcompartments regulates cytokine signaling

    Science Advances 2:e1600452.

    https://doi.org/10.1126/sciadv.1600452

    • Google Scholar
85. 1. Zhao W
    2. Lee C
    3. Piganis R
    4. Plumlee C
    5. de Weerd N
    6. Hertzog PJ
    7. Schindler C

    (2008) A conserved IFN-alpha receptor tyrosine motif directs the biological response to type I IFNs

    The Journal of Immunology 180:5483–5489.

    https://doi.org/10.4049/jimmunol.180.8.5483

    • PubMed
    • Google Scholar
86. 1. Zinkle A
    2. Mohammadi M

    (2018) A threshold model for receptor tyrosine kinase signaling specificity and cell fate determination

    F1000Research 7:872.

    https://doi.org/10.12688/f1000research.14143.1

    • Google Scholar

Acceptance summary:

This manuscript by Martinez-Fabregas et al. provides interesting evidence for a role of ligand affinity-dependent regulation of receptor trafficking leading to modification of intracellular signalling engagements and outcomes, using the IL6-IL6R system. It addresses a biologically important and technically interesting question of how binding affinity between cytokine ligands and receptors can differentially modulate signalling output. The case with cytokine receptors is contrasted with the biased signalling known in GPCRs, where different ligands produce different responses from the same receptor via allosteric modulations. To study this system, the authors generated a panel of IL-6 variants with differing binding affinities for gp130. They find that these ligands trigger biased signalling based on their affinities and use this system to extract mechanistic insights. More broadly, the authors suggest these results highlight that manipulation of cytokine-receptor binding parameters could be a useful strategy to decouple cytokine functional pleiotropy, which is a source of toxicity in cytokine-based therapies. This is a very impressive piece of work and an important contribution to the field.

Decision letter after peer review:

Thank you for sending your article entitled "Kinetics of cytokine receptor trafficking determine signaling and functional selectivity" for peer review at eLife. Your article is being evaluated by two peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation is being overseen by Satyajit Rath as the Senior Editor.

Given the list of essential revisions, including new experiments, the editors and reviewers invite you to respond within the next two weeks with an action plan and timetable for the completion of the additional work. We plan to share your responses with the reviewers and then issue a binding recommendation.

Both reviewers have evaluated your manuscript and find that your work addresses a biologically important and technically interesting question of how binding affinity between cytokine ligands and receptors can differentially modulate signaling output. However, both feel that the mechanism that you propose is not clearly described and alternative possibilities are not convincingly excluded. The important points to address are: 1) the data suggests a kinetic effect rather than an affinity effect, whereas an affinity effect is proposed, and 2) the alternative explanation that the different affinity ligands bind in different configurations, leading to differential receptor clustering and thereby different signalling output should be disproven. Some of the imaging data or experiments that affect trafficking are not clear. The labelled ligand experiments with confocal microscopy should be performed as time lapse as function of different ligand concentrations to convincingly demonstrate the differential trafficking of the high/low affinity ligands. In the PITSTOP experiment no statistical significance is given and only again the high affinity HylL6 is showing a clear effect. Are there alternative ways (genetic?) to affect the trafficking that yield more convincing results. In addition, please address point-by-point the comments of the individual reviews.

Reviewer #1:

This study claims that ligand-receptor lifetimes (investigated by generating iL6 mutants with different binding affinities) affect cytokine signalling trough differential trafficking of multimeric receptors comprising gp130. IL-6 stimulates signaling via a hexameric receptor comprised of two copies of IL-6, two gp130 and two IL-6Rα receptor subunits. The authors used yeast surface display to engineer a series of IL-6 variants binding gp130 with different affinities. Three variants with different apparent affinities for the receptor complex were investigated: A1 (648 nM) and the high affinity C7 (6.2 nM) and Mut3 (379 pM). My initial enthusiasm for this study was hampered by inconsistencies in some of the interpretation of the experiments. The major point that the authors make is that low affine binding of ligand to the receptor complex impedes receptor internalization leading to STAT3 biased responses at the plasma membrane whereas high affine binding to gp130 causes its internalization also enabling the activation of STAT1 on endosomes. The difference in the ratio of STAT1/STAT3 activity causes differential gene expression. The authors explain this difference in trafficking by the difference in the receptor-ligand lifetime (effectively the difference in koff). In the first result chapter a minor inconsistency confused me: at the end of the paragraph "The binding affinities of three of those variants, belonging to the first and second libraries, were confirmed by surface plasmon resonance (SPR) studies, with values ranging from 648 nM (A1) to 6.2 nM (C7) and 379 pM (Mut3) (Figure 1G and Figure 1—figure supplement 1C). The table in Figure 1G confirms that both Mut3 and C7 are high affinity, low nM Kd ligands with a KD ratio of approximately 16 (6nM/0.376nM). However, in Figure 1E where gp130 binding is measured the affinities seem to differ by two orders of magnitude for C7 and Mut3, whereas A1 does not seem to bind at all (despite a having a similar affinity as IL-6). What is the binding affinity of HylL-6 for gp130? In general, several further inconsistencies weaken the conclusions drawn in this manuscript. For example, the high affinity C7 mutant ligand is two orders of magnitude less potent in activating STAT1/STAT3 as the high affinity ligand HyIL-6 (Figure 2A,B). Even for the sub-nano-molar Mut3 mutant (super)-high affinity binder, its potency is one order of magnitude less in activating STAT1/STAT3 as the high affinity ligand HyIL-6. Why is this so? Is the affinity of HylL-6 even an order of magnitude higher as that of Mut3? Is a possible explanation that the different ligands induce a different receptor complex configuration by changing the interactions between extracellular domains? Could the balance between IL-6Rα engagement versus gp130 engagement affect signaling outcome? This is indicated by the fact that IL-6 wild type (wt), requires IL-6Rα expression to activate signalling whereas the high affinity Hyper IL-6 (HyIL-6) only requires gp130 and potently triggers signaling in cells lacking IL-6Rα. This all points at a possible other mechanism, where the different mutant ligands generate different extracellular conformations of the gp130 that could engage in different heterotypic receptor clusters with different signalling capacities.

The imaging data that should support the hypothesis that the lifetime of the receptor-ligand complexes affects the trafficking are not clear. For example, the authors detected the high affinity Mut3 ligand in intracellular compartments that partially co-localized with EEA1 early endosome marker, whereas no fluorescence was detected for the high affinity C7 mutant ligand (factor 3 difference in off rate) (Figure 3D,E). Should this not give some internalisation for high affinity ligands? For any ligand at least binding at the plasma membrane should be visible for concentrations higher than the Kd of the ligand/receptor complex, which is not visible for C7 (40 nM ligand concentration versus 6 nM Kd; Figure 3D). The off-rates given in table one let us approximate the lifetime of the complexes which for the C7 and Mut3 is in the order of ~20 min and ~60 min, which is very long in comparison with the STAT signalling time-scale. The authors also state: "We detected significant ligand induced gp130 co-locomotion for all IL-6 variants except for A1, with levels of co-locomotion paralleling gp130 binding affinities, i.e. HyIL-6>Mut3>C7>A1. Should A1 not have similar to WT IL-6 some engagement of GP130? Is this also not dependent on the outside concentration of A1 (saturating concentrations)? In general, the co-locomotion of receptors for any of the mutant ligands is very sparse (Figure 3—figure supplement 2). What is the reason for this if the functional receptor is a hexameric complex? When the authors compare gp130 degradation, only the high affinity ligand HyIL-6 causes significant degradation. Is it not so that if Mut3 is the high affinity ligand that was actually found on endosomes (Figure 3D) that it should also end up in lysosomes by the usual Rab5/Rab7 switch for early to late endosomes? In the PITSTOP experiment no statistical significance is given and only again the high affinity HylL6 is showing a clear effect. However, this effect on receptor degradation is similar to the non-stimulated case, making the whole link to ligand-receptor lifetime regulating trafficking weaker. The idea of competitive binding for the STATs to gp130 should also be true for gp130 on intracellular compartments? This idea could also be further tested by enhancing the relative expression of STAT1 (instead of knock down of STAT3) versus STAT3 in cells. Is this effect not due to a differential regulation of STAT-binding phosphorylation sites on gp130 by intra-cellular protein tyrosine phosphatases such as PTPN1/2? The factor signalling time is also not taken into consideration in the down-stream differential gene expression effect. Despite the interesting results that are presented, I feel that at this stage the authors have to do substantial additional work to prove that ligand-receptor lifetimes affect receptor trafficking and thereby multimeric receptor signalling via STAT1/STAT3.

Reviewer #2:

This manuscript by Martinez-Fabregas et al. addresses a biologically important and technically interesting question of how binding affinity between cytokine ligands and receptors can differentially modulate signaling output. The authors nicely frame the biological background of this problem in the Introduction. The case with cytokine receptors is contrasted with the biased signaling known in GPCRs, where different ligands produce different responses from the same receptor via allosteric modulations. The author's point out that this cannot be applied to cytokine receptors with a single membrane-spanning domain (although I question their certainty on this, see minor points).

To study this system, the authors generated a panel of IL-6 variants with differing binding affinities for gp130. They find that these ligands trigger biased signaling based on their affinities and use this system to extract mechanistic insights. More broadly, the authors suggest these results highlight that manipulation of cytokine-receptor binding parameters could be a useful strategy to decouple cytokine functional pleiotropy, which is a source of toxicity in cytokine-based therapies. Overall I think it's a very impressive piece of work and an important contribution to the field. I list a couple of more detailed points and questions below:

• In their study of differential STAT3/STAT1 activation the authors note that the different ligands drove differential phosphorylation amplitudes, which could not be rescued by increases in ligand concentration. This observation suggests that it is not affinity to which the receptors respond, but rather must be kinetic off-rate. If increasing ligand concentration cannot rescue the signaling differences, then receptor occupancy is not what the cell is measuring. The only other parameter is binding dwell time-very much the way T cell receptor distinguishes peptide MHC ligands. I suggest that the author's data does not support binding affinity as the differentiator, but rather binding dwell time. In the language they use in the text, this distinction is important and it would be useful if the authors elaborated on it more explicitly.

• The next section, "short lived IL-6-gp130 complexes fail to traffic to intracellular compartments", does essentially claim a kinetic discrimination mechanism-although the authors don't directly say this. The language still uses terms like "binding energy". I think the authors should take a more clear stand on kinetic discrimination vs. binding strength. Statements such as "one of the first steps affected by changes in gp130 binding affinity would be the assembly kinetics" are rather unclear. Why would binding affinity affect this? Binding affinity only affects receptor occupancy, and should always be compensated by changing ligand concentration. If the data indicate ligand concentration cannot compensate, then a kinetic proof reading mechanism in which short dwelling events fail to activate-no matter how many there are-seems likely, if not necessary. The nuance here is that kinetic proof reading requires an out-of-equilibrium process and rate-limiting kinetic intermediates (see, for example, Proc. Natl. Acad. Sci. USA 2016, 113, 29: 8218 and Science 2019, 363: 1098). Is the gp130 system capable of kinetic proof reading based on what is known? Perhaps internalization breaks equilibrium and provides the rate-limiting step? It’s not obvious to me that it is, but being able to say anything about this based on their data would be extremely interesting.

• In the Discussion, the author's seem to dismiss kinetic proof reading as unlikely or difficult to reconcile. However, isn't the STAT competition model the authors articulate providing a form of kinetic proof reading? To some degree this is only a matter of semantics in the definition of the term, kinetic proof reading. But I suggest the fact that the mechanism the authors have identified achieves a form of kinetic discrimination is an important realization, and probably should be identified as a form of kinetic proof reading.

Both reviewers have evaluated your manuscript and find that your work addresses a biologically important and technically interesting question of how binding affinity between cytokine ligands and receptors can differentially modulate signaling output. However, both feel that the mechanism that you propose is not clearly described and alternative possibilities are not convincingly excluded.

The important points to address are: 1) the data suggests a kinetic effect rather than an affinity effect, whereas an affinity effect is proposed.

We fully agree with the reviewers that our data strongly argues in favor of a kinetic effect rather than affinity, and we use terms like complex half-life and kinetics proof-reading throughout the manuscript to convey this. However, we see now that message did not come across in the manuscript clearly enough and we apologize for any confusion resulting from our language. We have amended the manuscript to clearly state this fact in the revised version. We provide some further points regarding this in the point-by-point answers to the reviewers (below).

2) the alternative explanation that the different affinity ligands bind in different configurations, leading to differential receptor clustering and thereby different signalling output should be disproven.

Despite our best efforts we have not been able to obtain protein crystals of our mutants in complex with gp130 and we can, therefore, not unequivocally exclude changes in binding topology. However, while it is possible that the new IL-6 variants bind gp130 in alternative conformations, we believe this to be highly improbable for the following reasons:

a) When generating the new IL-6 variants we kept the backbone of the cytokine intact and introduced mutations only in the cytokine-receptor binding interface. There is a large body of work done by the Garcia lab and others showing that introducing mutations in cytokine-receptor binding interfaces is a valid strategy to fine-tune binding affinities without changing topology (Levin AM, Nature, 2012; Juntilla IS, Nat.Chem.Biol., 2012; Moraga et al., 2015; Mendoza et al., 2017). In fact, to our knowledge there is not a single example in the literature of a mutation in a cytokine binding interface that alters the overall binding architecture of the ligand-receptor complex, strongly suggesting that the evolutionary pressure that maintains the canonical architecture of the cytokine-receptor complex is too high to be overcome by introducing few mutations in the binding interface. Indeed, the only ligands reported to alter canonical binding architectures are surrogate ligands such as diabodies and peptides (Livnah et al., 1998; Moraga et al., 2015; Mohan K, Science, 2019).

b) We have additional data, which we can include upon request, showing that IL-6 variants still rely on a classical site-2/site-3 interaction to recruit gp130 and signal, strongly supporting the hypothesis that our mutants recruit gp130 in a similar manner to wild type IL-6. We have generated a dominant-negative version of Mut3, by introducing mutations that disrupt its binding to gp130 via site-3. As a consequence, this new Mut3 variant behaves as an antagonist, able to bind gp130 with high affinity via site-2, but not able to trigger signaling due to its inability to recruit a second molecule of gp130 via site-3. These data confirm that Mut3 uses a site-2/site-3 architecture to assemble a signaling-compatible gp130 complex.

c) We have performed new TIRF microscopy experiments to confirm the stoichiometry of the complex formed by the IL-6 mutants. These experiments have shown that all IL-6 mutants, as well as HyIL-6, recruit two molecules of gp130 to trigger signaling, excluding the possibility of larger clusters of receptors induced by the mutants. This new data are now part of a new Figure 3 and Figure 3—figure supplements 1 and 2.

We believe that we have strong experimental evidences that support a canonical gp130 binding architecture by the different IL-6 mutants. However, in the absence of a crystal structure we agree that we cannot formally disprove this possibility and we have reflected this in the Discussion.

Some of the imaging data or experiments that affect trafficking are not clear. The labelled ligand experiments with confocal microscopy should be performed as time lapse as function of different ligand concentrations to convincingly demonstrate the differential trafficking of the high/low affinity ligands.

We agree with the reviewers that this part of the manuscript needs to be strengthened. We have now developed a flow cytometry-based method to monitor IL-6 internalization at a wide range of time and ligand concentrations. These new data sets, presented now in new Figure 3F-G, further reinforce our initial observations that short-lived IL-6-gp130 complexes fail to traffic to intracellular compartments and to induce gp130 degradation.

In the PITSTOP experiment no statistical significance is given and only again the high affinity HylL6 is showing a clear effect.

Blocking receptor internalization is very challenging. High doses of inhibitors like Pitstop will stress the cells leading to overall poor signaling. Complete silencing of clathrin or the use of dynamin dominant negative mutants produces similar outcome. Based on this, we had to find a compromise between inhibition of traffic and cellular health, which led to variability in the experimental outcome. However, we obtained similar effects on signaling when using Pitstop or clathrin silencing, which we interpret as a solid evidence that gp130 traffic was required for STAT1 activation but not STAT3 activation. Nevertheless, we understand the reviewer's concerns and we have now included new data as new Figure 4 A-B, which more clearly show that IL-6 variants exhibit a significant defect in inducing gp130 degradation as compared to HyIL-6. Additionally, using fluorescently labeled HyIL-6, we have now more accurately quantified the inhibition of gp130 internalization upon Pitstop treatment (new Figure 4C). Indeed, Pitstop treatment elicits a very strong inhibition of HyIL-6 internalization. Below, we expand on why only HyIL-6 induces degradation of gp130 in the answers to the reviewer 1 questions.

Overall, we believe that the new modifications that we have introduced in the text to better explain our conclusions, together with the large array of new experimental data that we are providing, fully support our initial hypotheses and address the majority of concerns raised by the editors and reviewers.

Reviewer #1:

This study claims that ligand-receptor lifetimes (investigated by generating iL6 mutants with different binding affinities) affect cytokine signalling trough differential trafficking of multimeric receptors comprising gp130. IL-6 stimulates signaling via a hexameric receptor comprised of two copies of IL-6, two gp130 and two IL-6Rα receptor subunits. The authors used yeast surface display to engineer a series of IL-6 variants binding gp130 with different affinities. Three variants with different apparent affinities for the receptor complex were investigated: A1 (648 nM) and the high affinity C7 (6.2 nM) and Mut3 (379 pM). My initial enthusiasm for this study was hampered by inconsistencies in some of the interpretation of the experiments. The major point that the authors make is that low affine binding of ligand to the receptor complex impedes receptor internalization leading to STAT3 biased responses at the plasma membrane whereas high affine binding to gp130 causes its internalization also enabling the activation of STAT1 on endosomes. The difference in the ratio of STAT1/STAT3 activity causes differential gene expression. The authors explain this difference in trafficking by the difference in the receptor-ligand lifetime (effectively the difference in koff). In the first result chapter a minor inconsistency confused me: at the end of the paragraph "The binding affinities of three of those variants, belonging to the first and second libraries, were confirmed by surface plasmon resonance (SPR) studies, with values ranging from 648 nM (A1) to 6.2 nM (C7) and 379 pM (Mut3) (Figure 1G and Figure 1—figure supplement 1C). The table in Figure 1G confirms that both Mut3 and C7 are high affinity, low nM Kd ligands with a KD ratio of approximately 16 (6nM/0.376nM). However, in Figure 1E where gp130 binding is measured the affinities seem to differ by two orders of magnitude for C7 and Mut3, whereas A1 does not seem to bind at all (despite a having a similar affinity as IL-6).

We appreciate the reviewer's comments and apologize for any confusion resulting from the lack of clarity in our presentation. The binding curves presented in Figure 1E correspond to the on-yeast binding titrations. Although this experiment is very informative and useful to stratify the different mutants based on their relative binding affinities for further studies, it does not provide a K_D binding constant as the one obtained in the SPR measurements. Additionally, it cannot detect binding of ligands with relatively low binding affinities, as A1 and IL-6 wt, explaining why we do not get binding with these variants in these experimental set up. Therefore, it is not surprising to see differences in binding values from the two different set of experiments, with the SPR providing the most accurate measurement (Moraga et al., 2015). We have now amended the text (Paragraph three of the Results section) to clarify this point.

What is the binding affinity of HylL-6 for gp130? In general, several further inconsistencies weaken the conclusions drawn in this manuscript. For example, the high affinity C7 mutant ligand is two orders of magnitude less potent in activating STAT1/STAT3 as the high affinity ligand HyIL-6 (Figure 2A,B). Even for the sub-nano-molar Mut3 mutant (super)-high affinity binder, its potency is one order of magnitude less in activating STAT1/STAT3 as the high affinity ligand HyIL-6. Why is this so? Is the affinity of HylL-6 even an order of magnitude higher as that of Mut3?

We apologize for the missing information on how IL-6 assembles its surface receptor to trigger signaling, which we believe has led to the confusion. We appreciate the point the reviewer is trying to make and have now clarified our language in the revised manuscript (Paragraph three of the Results section). Additionally, we are now providing schematic diagrams displaying the IL-6 receptor assembly kinetics modes used by the different IL-6 ligands described in this study as a new Figure 1A in the revised manuscript.

Briefly, IL-6 in a first step binds IL-6Rα receptor with high affinity. In a second step, this binary complex recruits gp130, via site-2 on the ligand-receptor interface, to form a hetero-trimeric complex, which cannot trigger signaling. In a third step, two hetero-trimeric complexes come together, via site-3 interaction, to finally form the signaling-compatible hexameric complex. HyIL-6, which is a fusion protein comprised of IL-6Rα and IL-6, directly binds gp130 on site-2 and then recruits a second gp130 molecule via site-3 to trigger signaling. It is important to emphasize that the presence of IL-6Rα helps to stabilize the complex by contributing to the formation of the site-2 and site-3 binding interfaces (new Figure 1—figure supplement 1A). Therefore, in the absence of IL-6Rα, IL-6 wildtype cannot bind gp130 and trigger signaling.

To produce our IL-6 variants we have mutated the site-2 binding interface on IL-6, with all of the mutants exhibiting an unaltered wild-type site-3 interface. Consequently, the IL-6 variants binding affinities reported in the manuscript correspond exclusively to binding of the mutants to site-2 and not the overall affinity of the complex. What this means is that we have improved the ability of the mutants to find gp130 on the cell surface and therefore their efficiency in forming the intermediate IL-6/gp130 (site-2) heterodimeric complex in the absence of IL-6Rα. However, in order to trigger signaling this heterodimeric (mutant/gp130) complex still requires the recruitment of a second molecule of gp130 via site-3, which exhibit very low affinity, ultimately defining the final half-life of the complex formed. HyIL-6, which comprises IL-6Rα, forms a more stable site-3 binding interface, by stabilization of this binding interface via IL-6Rα/gp130 contacts, therefore forming a longer-lived complex than our engineered mutants, explaining its stronger overall potency.

We have experimental data that support this binding sequence for the engineered variants, using a Mut3 dominant negative form and can provided upon request. We have ablated site-3 binding on Mut3 by replacing W157, which is located in the center of this binding interface, with a cysteine residue, allowing us to place a FITC molecule via a click chemistry reaction. This in turn completely ablates the ability of Mut3 to recruit a second gp130 molecule via its site-3 interface to trigger signaling. Importantly, Mut3DN inhibits signaling by the lower affinity C7 mutant proving that it is still binding with high affinity to gp130 via its site-2 interface. However, Mut3DN cannot inhibit signaling induced by HyIL-6, even at 50-fold excess, strongly supporting a more stable complex formed by HyIL-6.

Is a possible explanation that the different ligands induce a different receptor complex configuration by changing the interactions between extracellular domains?

Although possible, we believe this possibility to be very unlikely. Please see the answer that we provide above.

Could the balance between IL-6Rα engagement versus gp130 engagement affect signaling outcome? This is indicated by the fact that IL-6 wild type (wt), requires IL-6Rα expression to activate signalling whereas the high affinity Hyper IL-6 (HyIL-6) only requires gp130 and potently triggers signaling in cells lacking IL-6Rα. This all points at a possible other mechanism, where the different mutant ligands generate different extracellular conformations of the gp130 that could engage in different heterotypic receptor clusters with different signalling capacities.

As described above IL-6Rα contributes to stabilize the IL-6/receptor complex by interacting with gp130 via site-2 and site-3 binding interfaces. Thus, in the absence of IL-6Rα, the wildtype cytokine cannot recruit gp130 due to its low affinity. HyIL6, which already has IL-6Rα as part of its core, can recruit gp130 and trigger signaling. By contrast, our mutants can bind gp130 via site-2 in the absence of IL-6Rα. While gp130 dimerization is mediated by interactions via site-3 (which does not differ between the different variants), the overall lifetime of the signaling complex for our engineered IL-6 (two IL-6 with two gp130 molecules) is probably limited by the site-2 interaction. We therefore expect differential complex stabilities which ultimately impact their ability to activate STAT1 and STAT3 proteins. We want to emphasize at this point that the signaling as well as the imaging studies presented in original Figure 2,3, 4 and 5 in the manuscript were done in cells lacking IL-6Ra expression in order to characterize the real contribution of gp130 binding to signaling output, therefore explaining the stronger potency exhibited by HyIL-6 and the lack of signaling by wild-type IL-6. To address the alternative scenario suggested by the reviewer we have now performed further TIRF microscopy assays. We now provide strong evidences that all mutants form a stoichiometrically identical complex as the one formed by IL-6 wildtype and HyIL-6, which argues against the formation of larger heterotypic complexes by these mutants (new Figure 3E and Figure 3—figure supplement 2D).

The imaging data that should support the hypothesis that the lifetime of the receptor-ligand complexes affects the trafficking are not clear. For example, the authors detected the high affinity Mut3 ligand in intracellular compartments that partially co-localized with EEA1 early endosome marker, whereas no fluorescence was detected for the high affinity C7 mutant ligand (factor 3 difference in off rate) (Figure 3D,E). Should this not give some internalisation for high affinity ligands? For any ligand at least binding at the plasma membrane should be visible for concentrations higher than the Kd of the ligand/receptor complex, which is not visible for C7 (40 nM ligand concentration versus 6 nM Kd; Figure 3D). The off-rates given in table one let us approximate the lifetime of the complexes which for the C7 and Mut3 is in the order of ~20 min and ~60 min, which is very long in comparison with the STAT signalling time-scale.

The affinity of the mutants for gp130 exclusively represents their binding to site-2 and not the overall stability of the complex formed. So, while the C7 and Mut3 mutants, as highlighted by the reviewer, will remain bound to gp130 for ~20 min and ~60 min, this is not true for the signaling-compatible tetrameric complex of two IL-6 and two gp130 molecules formed by these mutants, which is limited by the affinity of the mutants for site-3. As a result, even Mut3, which binds with extremely high affinity to site-2, induces a very poor internalization when compared to hyIL-6, which exhibits high binding affinity for both site-2 and site-3 binding interfaces. This said, we agree with the reviewer in that the internalization experiment should be done at a wider range of ligand concentrations and stimulation times to more accurately assess gp130 internalization by the different IL-6 mutants. To address this, we have utilized a flow-cytometry based assay. We have incubated HeLa cells with a wide range of concentrations of the different fluorescently labelled IL-6 variants, or we saturated concentrations at different time points. We have used flow cytometry to quantify the fluorescence signals resulting from intracellular fluorescence molecules. We now show in new Figure 3F and G and 4C that HyIL-6 is the only IL-6 variant that it is robustly internalized in both experimental set ups, with the other IL-6 variants exhibiting very poor internalization, thus supporting our initial observations.

The authors also state: "We detected significant ligand induced gp130 co-locomotion for all IL-6 variants except for A1, with levels of co-locomotion paralleling gp130 binding affinities, i.e. HyIL-6>Mut3>C7>A1. Should A1 not have similar to WT IL-6 some engagement of GP130? Is this also not dependent on the outside concentration of A1 (saturating concentrations)? In general, the co-locomotion of receptors for any of the mutant ligands is very sparse (Figure 3—figure supplement 2). What is the reason for this if the functional receptor is a hexameric complex?

We agree with the reviewer in that the co-locomotion, although reproducible and significant, is sparse. The reason for that is that due to ubiquitous expression of endogenous gp130 in all cell lines, which affect TIRF measurements, we used an available gp130 KO HEK293 clone obtained by CRISPR technology (Schwerd T, J. Ex. Med., 2017). Unfortunately, HEK293 cells are very poor cells for imaging, therefore producing sparse signal as detected here. We have now generated new RPE1 gp130 KO clones, which we have used to repeat the TIRF studies. We are now providing new data shown in new Figure 3B and Figure 3—figure supplement 1 and 2, that robustly show dimerization efficiencies induced by the different IL-6 ligands. Additionally, we have performed new stoichiometric analysis to show that the IL-6 variants are not inducing different receptor oligomerization on the cell surface (new Figure 3E). It is notable that RPE1 cells do not express IL-6Rα, therefore we are not measuring hexameric complex formation, but tetrameric complex (two IL-6 with two gp130) for the different IL-6 variants and IL-6 wild-type, with the exception of HyIL-6, which already has IL-6Rα. This explains why we detect very low dimerization of gp130 in response to IL-6 wt or A1 mutant.

When the authors compare gp130 degradation, only the high affinity ligand HyIL-6 causes significant degradation. Is it not so that if Mut3 is the high affinity ligand that was actually found on endosomes (Figure 3D) that it should also end up in lysosomes by the usual Rab5/Rab7 switch for early to late endosomes? In the PITSTOP experiment no statistical significance is given and only again the high affinity HylL6 is showing a clear effect. However, this effect on receptor degradation is similar to the non-stimulated case, making the whole link to ligand-receptor lifetime regulating trafficking weaker.

HyIL-6 forms a more stable complex with gp130 than the IL-6 mutants due to its stronger interaction via site-3. The different mutants, due to their weak site-3 interaction, are unable to induce strong receptor internalization and therefore degradation. Although it is true that we detect some Mut3 molecules in endosomal compartments, the levels detected are much weaker than those induced by HyIL-6, and are insufficient to produce a significant decrease in the total levels of gp130. We have now performed more detailed kinetics of gp130 degradation (new Figure 4A and B) by the different IL-6 ligands to strengthen our initial observations. Our new data clearly show that HyIL-6 induces a significantly stronger degradation than the other IL-6 ligands, followed by the Mut3 mutant. C7, A1 and IL-6 ligands do not induce gp130 degradation, but lead to some stabilization of the receptor.

The idea of competitive binding for the STATs to gp130 should also be true for gp130 on intracellular compartments? This idea could also be further tested by enhancing the relative expression of STAT1 (instead of knock down of STAT3) versus STAT3 in cells.

We thank the reviewer for their suggestion. We now provide new data (new Figure 5F) where we show the effect of STAT1 overexpression in STAT3 phosphorylation by hyIL-6. Indeed, over expression of STAT1 results in decreased STAT3 phosphorylation induced by all IL-6 variants tested, further supporting our STAT competition model.

Is this effect not due to a differential regulation of STAT-binding phosphorylation sites on gp130 by intra-cellular protein tyrosine phosphatases such as PTPN1/2? The factor signalling time is also not taken into consideration in the down-stream differential gene expression effect. Despite the interesting results that are presented, I feel that at this stage he authors have to do substantial additional work to prove that ligand-receptor lifetimes affect receptor trafficking and thereby multimeric receptor signalling via STAT1/STAT3.

Although this is a possibility, we think that it is very unlikely. It has been previously reported that STAT1 and STAT3 compete for the same phospho-Tyr on gp130 (Gerhartz et al., 1996; Schmitz et al., 2000). However, we understand the reviewer point and now provide new data (new Figure 5C-D) where we have studied the contribution of each Tyr on the gp130 ICD to STAT1 and STAT3 phosphorylation upon HyIL-6 treatment. Our data clearly show that STAT1 and STAT3 compete for the same Tyr in gp130, and only after mutating all four Tyr to Phe we were able to block STAT1/STAT3 activation by HyIL-6.

Reviewer #2:

This manuscript by Martinez-Fabregas et al. addresses a biologically important and technically interesting question of how binding affinity between cytokine ligands and receptors can differentially modulate signaling output. The authors nicely frame the biological background of this problem in the Introduction. The case with cytokine receptors is contrasted with the biased signaling known in GPCRs, where different ligands produce different responses from the same receptor via allosteric modulations. The author's point out that this cannot be applied to cytokine receptors with a single membrane-spanning domain (although I question their certainty on this, see minor points).

To study this system, the authors generated a panel of IL-6 variants with differing binding affinities for gp130. They find that these ligands trigger biased signaling based on their affinities and use this system to extract mechanistic insights. More broadly, the authors suggest these results highlight that manipulation of cytokine-receptor binding parameters could be a useful strategy to decouple cytokine functional pleiotropy, which is a source of toxicity in cytokine-based therapies. Overall I think it's a very impressive piece of work and an important contribution to the field. I list a couple of more detailed points and questions below:

• In their study of differential STAT3/STAT1 activation the authors note that the different ligands drove differential phosphorylation amplitudes, which could not be rescued by increases in ligand concentration. This observation suggests that it is not affinity to which the receptors respond, but rather must be kinetic off-rate. If increasing ligand concentration cannot rescue the signaling differences, then receptor occupancy is not what the cell is measuring. The only other parameter is binding dwell time-very much the way T cell receptor distinguishes peptide MHC ligands. I suggest that the author's data does not support binding affinity as the differentiator, but rather binding dwell time. In the language they use in the text, this distinction is important and it would be useful if the authors elaborated on it more explicitly.

We fully agree with the reviewer. We have amended the text to explicitly state that IL-6/gp130 binding dwell time is the major factor contributing to biased signaling in the gp130 system.

• The next section, "short lived IL-6-gp130 complexes fail to traffic to intracellular compartments", does essentially claim a kinetic discrimination mechanism-although the authors don't directly say this. The language still uses terms like "binding energy". I think the authors should take a more clear stand on kinetic discrimination vs. binding strength. Statements such as "one of the first steps affected by changes in gp130 binding affinity would be the assembly kinetics" are rather unclear. Why would binding affinity affect this? Binding affinity only affects receptor occupancy, and should always be compensated by changing ligand concentration. If the data indicate ligand concentration cannot compensate, then a kinetic proof reading mechanism in which short dwelling events fail to activate-no matter how many there are-seems likely, if not necessary. The nuance here is that kinetic proof reading requires an out-of-equilibrium process and rate-limiting kinetic intermediates (see, for example, Proc. Natl. Acad. Sci. USA 2016, 113, 29: 8218 and Science 2019, 363: 1098). Is the gp130 system capable of kinetic proof reading based on what is known? Perhaps internalization breaks equilibrium and provides the rate-limiting step? It’s not obvious to me that it is, but being able to say anything about this based on their data would be extremely interesting.

We thank the reviewer for their suggestions. We agree that the most likely scenario that explains our results is a kinetics discrimination mechanism. Indeed, we had stated in the Discussion that “These data suggest that non-detectable short-lived IL-6/gp130 complexes can partially engage signaling, but fail to trigger a full response, evoking a kinetic-proof reading model”. However, we agree with the reviewer that the terminology should be more clear and possible mechanisms need to be discussed. We have amended the language in the Discussion to better define these concepts and more clearly state our conclusions. Regarding the reviewer's comments stating the need of an out-of-equilibrium process and rate-limiting kinetics intermediates for a kinetic-proof reading model, we believe that the gp130 system exhibits those requisites. As pointed out by the reviewer, receptor internalization is the critical event that breaks the equilibrium by irreversibly capturing intact signaling complexes into endosomes. We have recently quantitatively described this effect in a systematic study with engineered IL-13 variants (Moraga et al., 2015), showing the critical role of interaction rate constants in such a non-equilibrium process. For the engineered gp130 agonists, we clearly observed that endocytosis is altered, which presumably is caused by differential stability of ligand-receptor interaction (i.e. the off-rate). We have followed the reviewer’s advice and introduce a new chapter in the Discussion to describe these concepts (Discussion paragraph four).

• In the Discussion, the author's seem to dismiss kinetic proof reading as unlikely or difficult to reconcile. However, isn't the STAT competition model the authors articulate providing a form of kinetic proof reading? To some degree this is only a matter of semantics in the definition of the term, kinetic proof reading. But I suggest the fact that the mechanism the authors have identified achieves a form of kinetic discrimination is an important realization, and probably should be identified as a form of kinetic proof reading.

We fully agree with the idea that cytokine receptor complexes use a kinetic proof reading mechanism to define their signaling output and biological responses. We initially decided to be cautious in our use of the term as there might be some nuances between the cytokine system and other systems (e.g. TCR system) where kinetic proof reading has been proposed. However, we agree that our STAT competition model could be englobed within a more general kinetic proof reading mechanism. We have amended the Discussion to state this more clearly.

Author details

1. Jonathan Martinez-Fabregas

   Division of Cell Signaling and Immunology, School of Life Sciences, University of Dundee, Dundee, United Kingdom

   Contribution

   Conceptualization, Investigation, Methodology

   Contributed equally with

   Stephan Wilmes and Luopin Wang

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-5809-065X

2. Stephan Wilmes

   Division of Cell Signaling and Immunology, School of Life Sciences, University of Dundee, Dundee, United Kingdom

   Contribution

   Conceptualization, Investigation, Methodology

   Contributed equally with

   Jonathan Martinez-Fabregas and Luopin Wang

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-4112-710X

3. Luopin Wang

   Department Computer Science, Purdue University, West Lafayette, United States

   Contribution

   Data curation, Formal analysis, Investigation

   Contributed equally with

   Jonathan Martinez-Fabregas and Stephan Wilmes

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-5758-4092

4. Maximillian Hafer

   Department of Biology, University of Osnabrück, Osnabrück, Germany

   Contribution

   Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0853-2637

5. Elizabeth Pohler

   Division of Cell Signaling and Immunology, School of Life Sciences, University of Dundee, Dundee, United Kingdom

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

6. Juliane Lokau

   Department of Pathology, Medical Faculty, Otto-von-Guericke-University Magdeburg, Magdeburg, Germany

   Contribution

   Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-2573-7282

7. Christoph Garbers

   Department of Pathology, Medical Faculty, Otto-von-Guericke-University Magdeburg, Magdeburg, Germany

   Contribution

   Conceptualization, Methodology

   Competing interests

   No competing interests declared

8. Adeline Cozzani

   INSERM UMR-S-11721, Centre de Recherche Jean-Pierre Aubert (JPARC), Institut pour la Recherche sur le Cancer de Lille (IRCL), Université de Lille, Lille, France

   Contribution

   Methodology

   Competing interests

   No competing interests declared

9. Paul K Fyfe

   Division of Cell Signaling and Immunology, School of Life Sciences, University of Dundee, Dundee, United Kingdom

   Contribution

   Resources

   Competing interests

   No competing interests declared

10. Jacob Piehler

    Department of Biology, University of Osnabrück, Osnabrück, Germany

    Contribution

    Conceptualization, Methodology

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-2143-2270

11. Majid Kazemian

    Department Computer Science, Purdue University, West Lafayette, United States

    Contribution

    Conceptualization, Formal analysis, Investigation, Methodology

    Contributed equally with

    Suman Mitra and Ignacio Moraga

    For correspondence

    kazemian@purdue.edu

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-7080-8820

12. Suman Mitra

    INSERM UMR-S-11721, Centre de Recherche Jean-Pierre Aubert (JPARC), Institut pour la Recherche sur le Cancer de Lille (IRCL), Université de Lille, Lille, France

    Contribution

    Conceptualization, Investigation, Methodology

    Contributed equally with

    Majid Kazemian and Ignacio Moraga

    For correspondence

    suman.mitra@inserm.fr

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-3426-371X

13. Ignacio Moraga

    Division of Cell Signaling and Immunology, School of Life Sciences, University of Dundee, Dundee, United Kingdom

    Contribution

    Conceptualization, Funding acquisition, Investigation, Methodology, Project administration

    Contributed equally with

    Majid Kazemian and Suman Mitra

    For correspondence

    IMoragagonzalez@dundee.ac.uk

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-9909-0701

• 3,254

  Page views

• 500

  Downloads

• 15

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.