Superior Treg-Expanding Properties of a Novel Dual-Acting Cytokine Fusion Protein

Autoimmune diseases are caused by uncontrolled endogenous immune responses against healthy cells. They may develop due to an impaired function of regulatory T cells (Tregs), which normally suppress self-specific effector immune cells. Interleukin 2 (IL-2) and tumor necrosis factor (TNF) have been identified as key players that promote expansion, function, and stability of Tregs. In vivo, both low-dose IL-2 therapy and TNF receptor 2 (TNFR2) agonism were shown to expand Tregs and alleviate autoimmunity. We here designed a novel dimeric dual-acting fusion cytokine, where mouse IL-2 is genetically linked to a TNFR2-selective single-chain TNF mutein (IL2-EHD2-sc-mTNFR2). IL2-EHD2-sc-mTNFR2 showed high affinity to TNFR2 and efficiently activated IL-2 and TNFR2-selective signaling pathways. Further, IL2-EHD2-sc-mTNFR2 promoted superior Treg expansion, with both the IL-2 and the TNFR2 agonist (sc-mTNFR2) component necessary for this biological response. Ultimately, we propose that IL2-EHD2-sc-mTNFR2 is a dual-acting cytokine that efficiently promotes Treg expansion and might have a superior therapeutic window than conventional IL-2-based drugs.


INTRODUCTION
Autoimmune diseases are typically caused by endogenous immune responses against healthy cells. To avoid immune responses against self-antigens, most self-reactive T cells are eliminated during T cell development in the thymus. During this selection process potentially harmful, autoreactive T cells with effector activity are released from the thymus. However, these self-specific T cells do usually not induce autoimmune diseases because, in a healthy immune system, they are suppressed in the periphery by regulatory T cells (Tregs) (Sakaguchi et al., 2012;Klatzmann and Abbas, 2015). Most autoimmune diseases display defects in either the number or function of Tregs in the peripheral blood or the site of inflammation (Dominguez-Villar and Hafler, 2018). Further, genetic or pharmacological disruption of Treg-mediated immune control resulted in development of autoimmune disease in humans and mice (Klatzmann and Abbas, 2015). Therefore, strategies to restore and/or increase Treg function are currently developed and evaluated for clinical use.
It was well established that the cytokine interleukin-2 (IL-2) is an autocrine survival and proliferation signal for T cells in general, and considered to be the major differentiation signal for naive T cells into effector T cells (Robb, 1984). However, subsequent studies showed that the dominant role of IL-2 is the maintenance of Treg cells (Nelson, 2004), and not the development of effector and memory T cells, since loss of IL-2 signaling led to a break-down of immune tolerance and homeostasis rather than to immunodeficiency (Klatzmann and Abbas, 2015). This was confirmed by studies showing that IL-2 is a key mediator of Treg differentiation, survival and function (Liao et al., 2013;Klatzmann and Abbas, 2015). IL-2 binds to the IL-2 receptor (IL-2R), which is generated by different combinations of three polypeptide chains, IL-2Ra (CD25), IL-2Rb (CD122), and IL-2Rg (CD132). Whereas CD25 alone binds IL-2 with low affinity, the combination of CD122 and CD132 together form a complex that binds IL-2 with intermediate affinity, and the combination of all three polypeptide chains forms the high-affinity IL-2R. The high-affinity trimeric IL-2R is constitutively expressed by Tregs, whereas other subpopulations of T cells only acquire CD25 upon activation (Liao et al., 2013).
Due to the high affinity of the trimeric IL-2R on Tregs they respond to 10-100-fold lower concentrations of IL-2 than T effector cells or other low/intermediate-affinity IL-2R expressing cells (Klatzmann and Abbas, 2015). Therefore, low-dose IL-2 therapies, which allow to expand Tregs while avoiding IL-2induced clonal expansion or activation of T effector cells, are currently under clinical investigation for the treatment of autoimmune diseases.
Next to IL-2, TNFR2 has been recently identified as an important regulator of Treg function. TNFR2 promotes expansion of mouse  and human (Chen et al., 2010) Tregs and stabilizes Tregs in an inflammatory environment (Chen et al., 2013). Interestingly, expression of TNFR2 defines a maximally suppressive subset of mouse Tregs (Chen et al., 2008), indicating that TNFR2 activation regulates suppressive activity of Tregs. Indeed, it was shown that exogenous activation of TNFR2 leads to expansion of Tregs, alleviates arthritic disease and protects from acute graft versus host disease (Chopra et al., 2016;Fischer et al., 2018;Lamontain et al., 2018). Our published data indicate that presence of IL-2 dramatically increases TNFR2-dependent Treg expansion in vitro (Fischer et al., 2017;Fischer et al., 2018). Because of the apparent requirement of a combined activation of IL-2 and TNFR2 signal pathways we have developed a novel, dualacting cytokine fusion protein, where IL-2 is fused by genetic engineering to a TNFR2-selective TNF mutein. We here present data showing that this fusion protein is efficiently activating both IL-2 and TNFR2 signaling pathways and which leads to superior expansion of Tregs.

Production and Purification of Fusion Proteins
Recombinant proteins were produced and purified as described previously (Dong et al., 2016). Briefly, HEK293-6E cells were grown in F17 medium (Life Technologies, Darmstadt, Germany), and transiently transfected with expression constructs of TNF muteins using polyethyleneimine (Sigma). On the next day, Tryptone N1 (Organotechnie, TekniScience, Terrebonne, Canada) was added to the cell culture and cells were cultivated for additional 4 days. Then, supernatant was collected, and recombinant proteins were purified by immobilized metal ion chromatography (IMAC). For this purpose, supernatant was loaded onto a column containing Ni-NTA agarose (Macherey-Nagel, Düren, Germany) and unbound proteins were washed away using IMAC wash buffer (50 mM sodium-phosphate-buffer). Bound proteins were eluted with IMAC elution buffer (50 mM sodium-phosphate-buffer, 250 mM imidazole) and dialyzed (cut-off 4-6 kDa, Roth, Karlsruhe, Germany) against PBS overnight at 4°C. Finally, eluted proteins were purified by size exclusion chromatography (SEC). Protein concentration was determined by measuring the absorbance at 280 nm. For Coomassie staining, 2 µg of the purified fusion proteins were denatured in Laemmli buffer, resolved by 8% SDS-PAGE and stained with Coomassie.

Size-Exclusion Chromatography
Approx. 20 µg protein was applied to a BioSep-SEC-S2000 column (Phenomenex, Aschaffenburg, Germany) equilibrated with PBS and eluted at a flow rate of 0.5 ml/min using highperformance liquid chromatography (HPLC). For determining the size of recombinant proteins, standard proteins were run under the same conditions.

TNFR-Binding Assay
ELISA plates (Greiner, Frickenhausen, Germany) were coated with mouse TNFR1-Fc or TNFR2-Fc fusion proteins at 1 µg/ml in PBS and incubated at 4°C overnight. Residual binding sites were blocked with 2% skim milk powder in PBS at RT for 2 h. TNF muteins were diluted in 2% skim milk powder in PBS and incubated for 1 h at RT. Bound proteins were detected with mouse monoclonal antibodies to TNF (clone HP8001; 1 µg/ml) or his-tag (Dianova; incubation for 1 h at RT) and HRP-conjugated anti-rabbit or anti-mouse IgG antibodies (diluted 1:10,000; incubation for 1 h at RT), followed by incubation with TMB substrate solution. Reaction was stopped by addition of 1 M H 2 SO 4 and the absorbance at 450 nm was determined with an absorbance reader (Multiskan FC, Thermo Scientific, Karlsruhe, Germany) and data were analyzed using the software Microsoft Excel and GraphPad Prism 4 (GraphPad, La Jolla, CA). Between each step, nonbound proteins were removed by washing 4 times with 0.005% Tween-20 in PBS.

Quartz Crystal Microbalance
Affinities of the fusion proteins for TNFR2 were determined by quartz crystal microbalance measurements (Attana Cell 200, Attana, Stockholm, Sweden). Therefore, TNFR2-Fc fusion proteins were chemically immobilized on a carboxyl sensor chip according to the manufacturer's protocol at a high density (270 Hz). Binding experiments were performed in PBST (PBS, 0.1% Tween 20) pH 7.4 at a flow rate of 25 ml/min at 37°C. The chip was regenerated with 10 mM glycine HCl pH 2.0. Before each measurement, a baseline was measured which was subtracted from the binding curve. Data were collected using Attaché Office software (Attana, Stockholm, Sweden) and TraceDrawer (Ridgview Instruments, Vange, Sweden). K d values were calculated from the measured association-and dissociation-rates (Kd = k o ff /k o n ) assuming a 1:1 binding stoichiometry.

Cytotoxicity Assay
L929 cells (1.5 x 10 4 cells/well) were grown in 96-well flat bottom cell culture plates overnight. L929 cells were treated with actinomycin D (1 µg/ml) for 30 minutes prior to addition of fusion proteins. Then, cells were incubated with different concentrations of the fusion proteins for 24 h at 37°C. Cells were washed with PBS and incubated with crystal violet (20% methanol; 0.5% crystal violet) for 20 minutes to stain viable cells. The dye was washed away under rinsing water and cells were airdried. Crystal violet was resolved with methanol and the optical density at 550 nm was determined. Each sample was analyzed in triplicates and data were analyzed using the software Microsoft Excel and GraphPad Prism (GraphPad, La Jolla, CA).

Enzyme-Linked Immunosorbent Assay
BV-2 cells were stimulated as indicated, supernatants were collected after 24 h and analyzed by an ELISA specific for Cxcl-2 (BV-2, R&D Systems, Minneapolis MN) according to the instructions of the manufacturer.
For determination of the inflammatory marker CRP in the blood, whole blood was withdrawn and analyzed by an ELISA specific for mouse CRP (R&D Systems, Minneapolis MN) according to the instructions of the manufacturer. IFNg and IL-6 levels in the blood were determined by ELISAs specific for mouse IFNg and IL-6 (Biolegend, Koblenz, Germany). The absorbance at 450 nm was determined and the amount of released Cxcl-2 or IL-6 was determined with the provided standard and calculated using the software GraphPad Prism.

IL-2-Dependent Proliferation Assay
CT6 cells were cultivated with IL-2 (10 U/ml) for 2 weeks. Then, cells were harvested by centrifugation (300 g, 5 min) and washed twice with PBS (300 g, 5 min) to remove residual IL-2. To measure IL-2 induced proliferation, cells (1.5 x 10 4 cells/well) were seeded in 96-well U form plates and incubated in presence of different concentrations of the fusion proteins for 48 h. Number of cells was determined by MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide) assay. Therefore, cells were incubated with MTT (0.5 mg/ml) for 2 h at 37°C. Then, lysis buffer (10% SDS, 20 nM HCl) was added, cells were lyzed overnight and optical density at 550 nm was determined. Each sample was analyzed in triplicates and data were analyzed using the software Microsoft Excel and GraphPad Prism.

Mouse T Cell Isolation and Culture
Spleens from C57BL/6 wildtype mice were dissociated through a 40 µm cell strainer and collected in 10 ml MACS buffer (PBS, 0.5% BSA, 2 mM EDTA). Splenocytes were centrifuged (300 g, 5 min) and washed once with 10 ml MACS buffer. Then CD3 + (purity > 98%) and CD4 + (purity > 96%) T cells were isolated using the FACS Aria III plated in aCD3-coated (6 h at 4°C) 96well (U form) plates and cultivated in presence of soluble aCD28 and the fusion proteins. If indicated in the figure legend, cells were incubated with inhibitors for IL2R or TNFR2 before addition of fusion proteins. After 4 days, expression of IFNg, CD25, and FoxP3 were determined by flow cytometry. For IFNg staining, cells were reactivated with PMA/ionomycin in presence of brefeldin A for 4 h. For cytoplasmatic staining, the FoxP3 Staining Buffer Set (Miltenyi Biotech; Bergisch-Gladbach, Germany) was used according to the manufacturer's instructions. Data were acquired using a MACSQuant Analyzer 10 (Miltenyi) and analyzed with FlowJo (FlowJo, LLC).

RNA Isolation, cDNA Synthesis, and Quantitative Real-Time PCR
Total RNA was extracted using the PureLink RNA Mini Kit (ThermoFisher) and then transcribed into cDNA with the Maxima First Strand cDNA Synthesis Kit for RT-qPCR (ThermoFisher). The obtained cDNA was used to determine the FOXP3 gene expression by quantitative real-time PCR (qPCR; CFX96, Biorad) using specific primers (FoxP3 fwd: 5'-G C C A T G G C A A T A G T T C C T T C -3 ' ; FoxP3 rev: 5' -CGAACATGCGAGTAAACCAA-3', GAPDH fwd: 5'-GTGGCAAAGTGGAGATTGTTGCC-3'; FoxP3 rev: 5'-GATGATGACCCGTTTGGCTCC-3') and the Brilliant III Ultra-Fast SYBR Green QPCR Master Mix (Agilent, Waldbronn, Germany). To determine the expression of distinct genes, DDCt values were determined by correlating the obtained values to the house-keeping gene GAPDH. Data are presented relative to the PBS controls as normalized fold expression.

Statistics
Data are presented as mean ± standard error of the mean (SEM) of n independent experiments or animals. Normal distribution was analyzed by Shapiro-Wilk normality test. Statistical analyses were performed by Student's t-test or analysis of variance, followed by a post hoc Tukey's range test. A value of p < 0.05 was considered statistically significant.
Binding of the fusion proteins to TNFR1 and TNFR2 was analyzed by ELISA using immobilized mouse TNFR1-Fc (mTNFR1-Fc) and mTNFR2-Fc fusion proteins. No binding to TNFR1 by EHD2-sc-mTNF R2 and IL2-EHD2 was observed at any tested concentration of the fusion proteins ( Figure 2A). Unexpectedly, higher concentrations (above 1 nM) of IL2-EHD2-sc-mTNF R2 showed marginal interactions with the TNFR1-Fc coating ( Figure 2A). However, the observed apparent binding at higher concentrations is likely irrelevant, since no bioactivity on TNFR1 was observed and the measurable bioactivity on TNFR2 was in the sub-nanomolar range ( Figure  2). Both EHD2-sc-mTNF R2 (EC 50 value of 0.031 nM) and IL2-EHD2-sc-mTNF R2 (EC 50 value of 0.053 nM) bound to mTNFR2-Fc with similar binding efficacy ( Figure 2A). As expected, IL2-EHD2 did not interact with TNFR1 or TNFR2 ( Figure 2A). Next, we used quartz crystal microbalance measurements to determine the affinity of the fusion proteins to mTNFR2-Fc in a real-time binding analysis. Using a highdensity chip (270 Hz) with saturated, immobilized TNFR2, we determined sub-nanomolar binding affinities of EHD2-sc-mTNF R2 and IL2-EHD2-sc-mTNF R2 ( Figure 2B). IL2-EHD2sc-mTNFR2 revealed a dissociation constant (K d ) of 0.89 nM with an association rate constant k on of 5.89×10 5 M −1 s −1 and a dissociation rate constant k off of 4.77×10 -4 M −1 s −1 . The K d of EHD2-sc-mTNF R2 was approximately 7-times lower (0.12 nM) with a k on of 1.2×10 6 M −1 s −1 and a k off of 1.41×10 -4 M −1 s −1 , indicating more stable binding of EHD2-sc-mTNF R2 than IL2-EHD2-sc-mTNF R2 to TNFR2. TNFR2-selectivity of the fusion proteins was confirmed using L929 cells. In contrast to wild-type soluble TNF (sTNF), which activates TNFR1 and induces cell death in L929 cells, none of the TNFR2-selective fusion proteins induced cell death in L929 cells ( Figure 2C), verifying that fusion of IL2 to EHD2-sc-mTNF R2 did not impact TNFR2 selectivity of this novel molecule.
To determine whether IL2-EHD2-sc-mTNF R2 may dosedependent impact Tregs and T effector cells differently, we determined the fraction of CD25 + activated T cells and CD25 + FoxP3 + Tregs after cultivation in presence of 0.01 nM, 0.1 nM, and 1 nM IL2-EHD2 or IL2-EHD2-sc-mTNF R2 . Whereas IL2-EHD2-sc-mTNF R2 expanded Tregs more efficiently at all tested concentrations, the fraction of activated CD25 + T cells induced by IL2-EHD2 and IL2-EHD2-sc-mTNF R2 was similar at a concentration of 0.01 nM. In contrast, higher concentrations of IL2-EHD2-sc-mTNF R2 also induced T cell activation more efficiently ( Figure 3C). We then confirmed that IL2-EHD2-sc-mTNF R2 expanded Tregs using a gene expression analysis. To this end, T cells were cultivated in presence of 1 nM of the fusion proteins for 4 days and gene expression of the Treg transcription factor FoxP3 was determined by quantitative real-time PCR. As expected incubation with IL2-EHD2-sc-mTNF R2 resulted in the highest FOXP3 expression levels ( Figure 3D), confirming the superior Treg expanding activity. We next investigated the impact of the fusion proteins on T effector cells using expression of the Th1 cytokine IFNg as a read-out. Therefore, T cells were cultivated in presence of the fusion proteins for 4 days. Then, cells were reactivated using PMA/ionomycin in presence of brefeldin A and expression of IFNg in the FoxP3cell population was analyzed by flow cytometry. No differences in the expression of IFNg between IL2-EHD2 and IL2-EHD2-sc-mTNF R2 cultivated cells were observed ( Figure 3E).
Finally, we analyzed Treg expansion via IL2-EHD2-sc-mTNF R2 compared to combination of its building blocks, IL2-EHD2 and EHD2-sc-mTNF R2 . At a concentration of 1 nM, both IL2-EHD2-sc-mTNF R2 and IL2-EHD2/EHD2-sc-mTNF R2 treatment induced similar expansion of CD25 + FoxP3 + Tregs, indicating that the genetic fusion of the single components does not result in improved bioactivity compared to the combination treatment ( Figure 4C). We then compared the impact of 0.01 nM and 1 nM fusion protein on proliferation of FoxP3 + Tregs and FoxP3 -T cells. To this end, we calculated the Ki-67 expression at a concentration of 0.01 nM relative to the expression at 1 nM of the fusion proteins ( Figure 4D). No differences in the expression of Ki-67 within the FoxP3 + Treg population were observed after stimulation with 0.01 nM and 1 nM of IL2-EHD2-sc-mTNF R2 and the combination of IL2-EHD2 and EHD2-sc-mTNF R2 . Similar, stimulation with 0.01 nM or 1 nM IL2-EHD2 and EHD2-sc-mTNF R2 resulted in a comparable Ki-67 expression in the FoxP3 -T cell subpopulation. In contrast, 0.01 nM of IL2-EHD2-sc-mTNF R2 induced a significantly reduced Ki-67 expression comparted to a higher concentration of 1 nM. This suggests that while higher concentrations of IL2-EHD2-sc-mTNF R2 or the combination of IL2-EHD2 and EHD2-sc-mTNF R2 induce proliferation of FoxP3 -T effector cells with a similar efficacy, lower concentrations of IL2-EHD2-sc-mTNF R2 loose their activity on FoxP3 -T effector cells, while the combination of IL2-EHD2 and EHD2-sc-mTNF R2 still impact T effector cell proliferation like higher concentrations.

DISCUSSION
We and others have shown that exogenous TNFR2 activation promotes the expansion of Tregs in vitro and in vivo (Okubo et al., 2013;Chopra et al., 2016;Okubo et al., 2016;Fischer et al., 2017;Fischer et al., 2018;Lamontain et al., 2018) and that TNFR2 and IL-2 synergistically promote Treg expansion (Fischer et al., 2018). We therefore genetically fused mouse IL-2 to a TNFR2-selective TNF mutein resulting in IL2-EHD2-sc-mTNF R2 , a fusion protein that efficiently activated both IL-2 and TNFR2 signaling. Our data show that IL2-EHD2-sc-mTNF R2 retained an affinity for TNFR2 in the subnanomolar range and efficiently activated TNFR2 and IL-2 signaling. IL2-EHD2-sc-mTNF R2 showed superior Tregexpanding activities compared to IL2-EHD2 and EHD2-sc-mTNF R2 stimulation alone, with both the IL2 and the EHD2sc-mTNF R2 component necessary for the increased bioactivity. Altogether, our data indicate that IL2-EHD2-sc-mTNF R2 is a promising novel dual-acting cytokine fusion protein that might be superior to current Treg expanding strategies based on lowdose IL-2 monotherapy.
Tregs only constitute 1% to 2% of peripheral blood lymphocytes, but they are the master controllers of selftolerance to avoid autoimmunity and guarantee long-term immune homeostasis (Sakaguchi et al., 2008). Accordingly, deregulation of Treg numbers or function have been implicated in the pathology of a variety of diseases. The most prominent example is the IPEX (immuno-dysregulation, polyendocrinopathy, enteropathy, X-linked) syndrome, a disease that develops due to mutations in the FOXP3 gene leading to Treg defects and resulting in lethal multi-organ FIGURE 4 | Phenotype of IL2-EHD2-sc-mTNF R2 expanded Tregs. CD4 + T cells were isolated from mouse splenocytes and activated using plate-bound anti-CD3 (5 µg/ml) and soluble anti-CD28 (2 µg/ml). (A, B) Cells were cultivated in presence of 1 nM of the oligomerized muteins for 4 days. (A) Cells were first gated for CD25 + cells and then percentage of proliferating Ki-67 + cells with the FoxP3 + and FoxP3cell population was determined by flow cytometry. Shown is a representative experiment (left panel) and quantification from three independent experiments ± SEM (analyzed by one-way ANOVA). (B) Expression of CD45RA, CD62L, CTLA-4, GITR, and ICOS was quantified in the CD25 + FoxP3 + subpopulation by flow cytometry. Shown are representative histograms from 3 independent experiments. (C) Cells were cultivated in presence of 1 nM IL2-EHD2-sc-mTNF R2 or the combination of IL2-EHD2 and EHD2-sc-mTNF R2 for 4 days. Percentage of FoxP3 + Tregs with the population of CD25 + cells was determined by flow cytometry. Shown is a representative experiment (left panel) and quantification from three independent experiments ± SEM (analyzed by one-way ANOVA). (D) Cells were cultivated in presence of 0.01 nM or 1 nM IL2-EHD2-sc-mTNF R2 or the combination of IL2-EHD2 and EHD2-sc-mTNF R2 for 4 days. Expression of Ki-67 was analyzed by flow cytometry. Results are presented as Ki-67 expression at a concentration of 0.01 nM relative to the Ki-67 expression at a concentration of 1 nM (n = 3 ± SEM, t test). *p < 0.05, **p < 0.01, ns, not significant. inflammation and autoimmunity (Bacchetta et al., 2018). Clinical studies of patients further revealed that there is a defect in either the number or the function of Treg cells isolated from the peripheral blood in various autoimmune disorders, such as type 1 diabetes, multiple sclerosis, systemic lupus erythematosus, myasthenia gravis and, rheumatoid arthritis (Miyara et al., 2011;Dominguez-Villar and Hafler, 2018). Thus, a number of ongoing clinical trials are repurposing approved drugs, such as the Treg growth factor IL-2 (aldesleukin) or Treg-stabilizing factors, e.g., the mTORinhibitor rapamycin to enhance Treg function as a therapeutic approach to control a variety of autoimmune diseases, GvHD, or organ transplant rejection (Abbas et al., 2018).
IL-2, which was first discovered as an autocrine growth factor for cultured T effector cells, also plays an important role for Treg survival and is essential for the functional capacity of Treg cells. IL-2 is required for the stable expression of Foxp3 (Fontenot et al., 2005;Feng et al., 2014) and other mediators of Treg cell-suppressive activity such as CD25 and CTLA-4 ( Barron et al., 2010). Due to its parallel role for T effector cell function, administration of high concentrations of IL-2 cause toxicity, including vascular leak syndrome and other manifestations of a cytokine storm (Abbas et al., 2018). It is now well established that low-dose IL-2 therapy preferentially activates Tregs because of the constitutive high expression of CD25 as a subunit of the trimeric high affinity IL-2R (Klatzmann and Abbas, 2015;Abbas et al., 2018). Clinical trials have been conducted in hepatitis virus-associated vasculitis (Saadoun et al., 2011), chronic graft-versus-host disease (GVHD) (Koreth et al., 2011), systemic lupus erythematosus and type 1 diabetes (Klatzmann and Abbas, 2015). However, despite promising results, a continuing concern with this approach is that the therapeutic window for doses may be small, due to potential activation of the effector arm of the immune system, thus carrying the risk of exacerbating disease (Abbas et al., 2018). Our data indicate that IL2-EHD2-sc-mTNF R2 shows superior Tregexpanding bioactivity, compared to either IL2-EHD2 or EHD2sc-mTNF R2 single treatments. Therefore, lower doses of this dualacting cytokine may be therapeutic, without inducing toxic side effects, i.e., we hypothesize that IL2-EHD2-sc-mTNF R2 might have a larger therapeutic window than conventional low dose IL-2 protocols. However, at a first glance our in vitro data presented here suggest that costimulation with IL2-EHD2 and EHD2-sc-mTNF R2 together displays equal bioactivity as the fusion protein at concentrations of 1 nM. However, scrutinizing potential differences in dose response relationships between Treg and Teff, we noted that at low doses, the co-application of the single acting cytokines was superior in Teff expansion as compared to the dual action cytokine fusion protein. Therefore, we propose that the therapeutic window of the dual acting IL2-EHD2-sc-mTNF R2 fusion protein should be superior to treatment with IL-2 only or a combination of IL-2 with a monospecific TNFR2 agonist.
Other functions of IL-2 include its ability to inhibit the generation of proinflammatory T helper 17 cells (Laurence et al., 2007) and may synergize with its effects on Treg cells in the treatment of autoimmune and inflammatory diseases. Similar, next to its important role for expansion, function and stability of mouse and human Tregs Chen et al., 2010;Nagar et al., 2010;Chen et al., 2013;Okubo et al., 2013), cell-intrinsic TNFR2 signaling has been shown to impair Th17 differentiation by promoting IL-2 expression (Miller et al., 2015). Further, TNFR2 has a well-documented role for tissue regeneration, i.e., TNFR2 promotes remyelination (Arnett et al., 2001), pancreatic regeneration (Ban et al., 2008;Faustman and Davis, 2013), and cardioprotection . In particular, TNFR2 was shown to selectively promote death of autoreactive T cells in human diabetes (Ban et al., 2008;Okubo et al., 2016) and promoting Treg activity is a promising approach to treat type I diabetes (Visperas and Vignali, 2016). Therefore, combining the bioactivity of IL-2 and a TNFR2 agonists in one molecule may be superior to conventional Treg expanding therapies due to (1) more efficient Treg expansion and (2) promotion of additional beneficial responses, such as tissue regeneration via the TNFR2-selective component.
In addition, efforts were described to engineer pharmacologically superior and Treg-selective IL-2 variants, e.g. by increasing the affinity of IL-2 for the a-chain (CD25) of the high-affinity-IL-2Rabg (Rao et al., 2005) or by decreasing the cytokines affinity to the bchain in the intermediate affinity IL-2Rbg (Peterson et al., 2018). As a future perspective, the combination of such a genetically optimized human IL-2 mutein with a human TNFR2-selective mutein (Fischer et al., 2011;Dong et al., 2016) might be a promising novel dual-acting drug prototype to promote immune tolerance and tissue regeneration.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/ supplementary material.

AUTHOR CONTRIBUTIONS
TP, MS, CH, NP, and RF conducted experiments. JB, KP, and RK designed parts of the experimental work, analyzed the results and contributed to writing the manuscript. RF designed and supervised the project, analyzed the results and wrote the manuscript.

FUNDING
RF was supported by the Carl Zeiss Foundation (Az. 0563-2.8./ 508/2) and a DFG research fellowship (FI 2138/1-1). The fusion proteins described in this work can be obtained via a Material Transfer Agreement (MTA).

ACKNOWLEDGMENTS
We thank Alexandra Kraske and Beatrice Reiser for animal care.