Molecular reshaping of phage-displayed Interleukin-2 at beta chain receptor interface to obtain potent super-agonists with improved developability profiles

The dataset contains 19 original figures constructed with Graphpad (including the corresponding source data) presented in the above-referred manuscript.

Authors: Gertrudis Rojas1*, Ernesto Relova-Hernndez1, Annia Prez-Rivern1, Camila Castro-Martnez1, Osmany Diaz-Bravo1, Yanelys Cabrera Infante1, Tania Gmez1, Joaqun Solozbal1, Ana Beatriz DazBravo1, Maren Schubert2, Marlies Becker2, Beatriz Prez-Massn1, Dayana Prez-Martnez1, Rydell Alvarez-Arzola1, Osmany Guirola3, Glay Chinea3, Luis Graca4, Stefan Dbel2, Kalet Len1, Tania Carmenate1

Author affiliations:
1Center of Molecular Immunology, calle 216 esq 15, apartado 16040, Atabey, Playa, CP 11300, La Habana, Cuba
2Technische Universitt Braunschweig, Institute of Biochemistry, Biotechnology and Bioinformatics, Department of Biotechnology, Spielmannstrae 7, 38106 Braunschweig, Germany
3 Center of Genetic Engineering and Biotechnology, Ave 31 e/ 158 y 190, apartado 6162, Playa, CP 11300, La Habana, Cuba
4Instituto de Medicina Molecular Joo Lobo Antunes, Faculdade de Medicina da Universidade de Lisboa, Centro Acadmico de Medicina de Lisboa, Lisbon, Portugal

*Correspondence should be addressed to G.R. (grojas@cim.sld.cu)

The contents of each file are detailed below:
Figure 1a. Diversity of the soft-randomization human IL-2 library. The library was constructed by Kunkel mutagenesis on a human IL-2 gene-containing phagemid template, using two mutagenic spiked oligonucleotides which targeted beta interface segments 12-23 and 81-95. The composition of oligonucleotides was biased to the original wild-type sequence, but introduced limited variability at eight and seven codons of these segments, respectively. Library phages were panned on immobilized extracellular domain of human IL-2 receptor beta chain (three selection rounds). Deducing protein sequences of a sample of cytokine-displaying clones picked before and after selection revealed that the number of amino acid (aa) replacements per sequence increased upon selection.

Figure 1b. Beta chain reactivity of phage pools selected from a soft-randomization human IL-2 library. The library was constructed by Kunkel mutagenesis on a phagemid template containing human IL-2 gene, using two mutagenic spiked oligonucleotides which targeted beta interface segments 12-23 and 81-95. The composition of oligonucleotides was biased to the original wild-type sequence, but introduced limited variability at eight and seven codons of these segments, respectively. Library phages were panned on immobilized extracellular domain (ECD) of human IL-2 receptor beta chain (three selection rounds). Beta chain reactivity of phage pools obtained before and after selection rounds was evaluated by ELISA on polyvinyl chloride microtitration plates coated with beta chain ECD. Bound phages were detected with an anti-M13 antibody conjugated to horseradish peroxidase. Phage-displayed versions of human IL-2 and H9 superkine were used as controls. Binding specificity was assessed with a non-related coating protein (bovine serum albumin). Each phage sample was evaluated twice on each coating protein. Bars indicate the mean absorbance values of both determinations, while dots represent individual data points.

Figure 2e. Charge shift after phage selection from the secondary library S1. The library was constructed by Kunkel mutagenesis on a human IL-2 gene-containing phagemid template, using a mutagenic oligonucleotide that introduced full randomization at positions 81, 83, 84 and 87, and limited variability (Pro/Ala) at position 82. Library phages were panned on immobilized extracellular domain of human IL-2 receptor beta chain (three rounds). Plasmid samples of cytokine-displaying clones from the unselected library and of those obtained after panning were sequenced, and protein sequences were deduced. Analysis of the net charge in the segment 81-87 within both groups revealed a shift towards higher negative charges upon selection.

Figure 2f. Beta chain reactivity of phage selected from the secondary library S1. The library was constructed by Kunkel mutagenesis on a human IL-2 gene-containing phagemid template, using a mutagenic oligonucleotide that introduced full randomization at positions 81, 83, 84 and 87, and limited variability (Pro/Ala) at position 82. Library phages were panned on immobilized extracellular domain (ECD) of human IL-2 receptor beta chain (three rounds). Reactivity of phage pools obtained before and after selection rounds was evaluated by ELISA on polyvinyl chloride microtitration plates coated with beta chain ECD. Bound phages were detected with an anti-M13 antibody conjugated to horseradish peroxidase Phage-displayed hIL-2 and H9 superkine were used as controls. Binding specificity was assessed with a non-related coating protein (bovine serum albumin). Each phage sample was evaluated twice on each coating protein. Bars indicate the mean absorbance values of both determinations, while dots represent individual data points.

Figure 3g. Beta chain reactivity of phage clones selected from the secondary library S2. Positions 81, 83, 84 and 87 were totally randomized in S2 library and P82 was replaced by the mixture Pro/Ala. L80, L85, I86, I89, I92 and V93 were substituted by a mixture of hydrophobic residues (Ile, Leu, Met, Phe and Val). Library was constructed by Kunkel mutagenesis on a human IL-2 gene-containing phagemid template. Phages were panned on immobilized extracellular domain (ECD) of IL-2 receptor beta chain (three rounds). Thirteen different sequences were retrieved after selection. Reactivity of purified phage preparations from selected clones was evaluated by ELISA on polyvinyl chloride microtitration plates coated with either beta chain ECD or anti-c-myc tag 9E10 antibody. Bound phages were detected with an anti-M13 antibody conjugated to horseradish peroxidase. Relative reactivity was calculated as the ratio between signals obtained with beta chain and with 9E10. Phage-displayed hIL-2 and H9 superkine were used as controls. Bars represent relative reactivities.

Figure 3h. Charge shift after phage selection from the secondary library S2. Positions 81, 83, 84 and 87 were totally randomized in S2 library and P82 was replaced by the mixture Pro/Ala. L80, L85, I86, I89, I92 and V93 were substituted by a mixture of hydrophobic residues (Ile, Leu, Met, Phe and Val). Library was constructed by Kunkel mutagenesis on a human IL-2 gene-containing phagemid template. Phages were panned on immobilized extracellular domain of IL-2 receptor beta chain (three rounds). Thirteen different sequences were retrieved after selection. Plasmid samples of cytokine-displaying clones from the unselected library and of those obtained after panning were sequenced, and protein sequences were deduced. Analysis of the net charge in the segment 81-87 within both groups revealed a shift towards higher negative charges upon selection.

Figure 4a. Characterization of beta chain reactivity of recombinant proteins derived from S1 library screening. Recombinant proteins comprising the new human IL-2 variants designed upon S1 library screening, fused to a human IgG1 Fc region, were produced by transient transfection of HEK-293T cells adapted to grow in suspension, and purified by protein A affinity chromatography. Control fusion proteins derived from human IL-2 (K35E), from its single-mutated variant (R81D) and from the H9 superkine, were included. The reactivity of the purified proteins (31.25 ng/mL) was evaluated by ELISA on polyvinyl chloride microtitration plates coated with a recombinant fusion protein comprising human IL-2R beta subunit extracellular domain and mouse IgG2a Fc. Bound proteins were detected with an anti-human IgG antibody conjugated to horseradish peroxidase. Each protein sample was evaluated twice. Bars indicate the mean absorbance values of both determinations, while dots represent individual data points.

Figure 5i. Characterization of thermal stability of recombinant proteins derived from S2 library screening. Recombinant proteins comprising the new human IL-2 variants designed upon S2 library screening, fused to a human IgG1 Fc region, were produced by transient transfection of HEK-293T cells adapted to grow in suspension, and purified by protein A affinity chromatography. Stability of the fusion proteins under thermal stress was assessed by ELISA after incubating them at 50C. Purified proteins (15.6 ng/mL) were incubated on polyvinyl chloride microtitration plates coated with the recombinant fusion protein comprising IL-2 receptor beta extracellular doman fused to a mouse IgG2a Fc domain. Bound proteins were detected with an anti-human IgG antibody conjugated to horseradish peroxidase. The fusion protein derived from from H9 superkine was included as control. Untreated proteins were also evaluated and used as references to calculate the relative reactivity (%) of each heated sample.

Figure 6d. Beta chain reactivity of recombinant proteins derived from human IL-2 variants with mutations at position 92. A panel of recombinant proteins comprising human IL-2 (K35E) and H9 superkine variants with or without additional replacements at position 92, fused to a human IgG1 Fc region, was produced by transient transfection of HEK-293T cells adapted to grow in suspension. The proteins were purified by protein A affinity chromatography. The reactivity of purified proteins was titrated by ELISA on polyvinyl chloride microplates coated with a recombinant fusion protein comprising IL-2R beta subunit extracellular domain and mouse IgG2a Fc. Bound proteins were detected with an anti-human IgG antibody conjugated to horseradish peroxidase. Each protein concentration was evaluated twice. Lines connect the mean absorbance values at different concentrations of the same protein. Symbols represent individual data points.

Figure 8a. Capacity of IL-2-derived recombinant fusion proteins to induce CTLL-2 proliferation. Mouse CTLL-2 cells were grown during 48h in the presence of serial dilutions of recombinant Fc-fusion proteins comprising the new IL-2 beta super-binders derived from library screening. Reference proteins IL-2(K35E)/Fc and H9(K35E)/Fc were also included. Fc-fused IL-6 was used as negative control. Cells were stained with Alamar Blue dye and the proliferation curves were constructed by plotting the difference between absorbances at 540 and 620 nm as a function of fusion protein concentration. Two replicates of each protein concentration were analysed. Symbols represent the values of both determinations.

Figure 8b. Capacity of IL-2-derived recombinant fusion proteins to induce CD25-KO CTLL-2 proliferation. CTLL-2 cells transduced with lentiviral particles encoding a sgRNA suitable to ablate expression of IL-2receptor alpha subunit gen (CD25-KO CTLL-2 model) were grown during 48h in the presence of serial dilutions of recombinant Fc-fusion proteins comprising the new IL-2 beta super-binders derived from library screening. Reference proteins IL-2(K35E)/Fc and H9(K35E)/Fc were also included. Fc-fused IL-6 was used as negative control. Cells were stained with Alamar Blue dye and the proliferation curves were constructed by plotting the difference between absorbances at 540 and 620 nm as a function of fusion protein concentration. Two replicates of each protein concentration were analysed. Symbols represent the values of both determinations.

Figure 8c. Capacity of IL-2-derived recombinant fusion proteins to induce STAT5 phosphorylation in CTLL-2 cells. CTLL-2 cells were stimulated during 40 min with the set of recombinant Fc-fusion proteins comprising the new IL-2 beta super-binders derived from library screening, fixed, permeabilized and stained with an APC-labelled antibody against phosphorylated STAT5. The resulting fluorescence signals (determined by flow cytometry) measured signalling capacity of the proteins on the cells. Fc-fused IL-6 was used as negative control. Two replicates of each protein concentration were analysed. Symbols represent the values of both determinations.

Figure 8d. Capacity of IL-2-derived recombinant fusion proteins to induce STAT5 phosphorylation in CD25-KO CTLL-2 cells. CTLL-2 cells transduced with lentiviral particles encoding a sgRNA suitable to ablate expression of IL-2receptor alpha subunit gen (CD25-KO CTLL-2 model) were stimulated during 40 min with the set of recombinant Fc-fusion proteins comprising the new IL-2 beta super-binders derived from library screening, fixed, permeabilized and stained with an APC-labelled antibody against phosphorylated STAT5. Fc-fused IL-6 was used as negative control. The resulting fluorescence signals (determined by flow cytometry) measured signalling capacity of the proteins on the cells. Two replicates of each protein concentration were analysed. Symbols represent the values of both determinations.

Figure 9b. Increase in the spleen cell content induced by IL-2-derived fusion proteins. Four groups of five mice each received daily injections (during four days) of recombinant proteins comprising IL-2-derived beta super-binders fused to human Fc(LALA). Two additional groups were treated with similarly formatted proteins containing either wild-type human IL-2(K35E) or H9 superkine. Mice from a control group received PBS injections. All the animals were sacrificed at the fifth day, and their spleens were collected and macerated for cell count. Dots indicate spleen cell content of individual animals. Lines represent mean values and error bars indicate SD within each group of five animals receiving the same treatment.

Figure 9d. In vivo expansion of effector T cell population CD8+CD44hiCD122hi induced by IL-2-derived fusion proteins. Four groups of five mice each received daily injections (during four days) of recombinant proteins comprising IL-2-derived beta super-binders fused to human Fc(LALA). Two additional groups were treated with similarly formatted proteins containing either wild-type human IL-2(K35E) or H9 superkine. Mice from a control group received PBS injections. All the animals were sacrificed at the fifth day, and their spleens were collected and macerated for characterization by flow cytometry. Dots indicate total numbers of CD8+CD44hiCD122hi T cells in individual animals Lines represent mean values and error bars indicate SD within each group of five animals receiving the same treatment.

Figure 9f. In vivo induction of effector cell proliferation induced by IL-2-derived fusion proteins. Four groups of five mice each received daily injections (during four days) of recombinant proteins comprising IL-2-derived beta super-binders fused to human Fc(LALA). Two additional groups were treated with similarly formatted proteins containing either human IL-2 (K35E) or H9 superkine. Mice from a control group received PBS injections. All the animals were sacrificed at the fifth day, and their spleens were collected and macerated for cell characterization by flow cytometry. Dots indicate the number of proliferating effector cells (CD8+Ki67+) of individual animals. Lines represent mean values and error bars indicate SD within each group of five animals receiving the same treatment.

Figure 9h. In vivo expansion of T regulatory cells induced by IL-2-derived fusion proteins. Four groups of five mice each received daily injections (during four days) of recombinant proteins comprising IL-2-derived beta super-binders fused to human Fc(LALA). Two additional groups were treated with similarly formatted proteins containing either human IL-2 (K35E) or H9 superkine. Mice from a control group received PBS injections. All the animals were sacrificed at the fifth day, and their spleens were collected and macerated for characterization by flow cytometry. Dots indicate the total numbers of Treg cells (CD4+CD25+Foxp3+) of each individual animal. Lines represent mean values and error bars indicate SD within each group of five animals receiving the same treatment.

Figure 10b. Figure 10. Anti-metastatic effects of IL-2-derived fusion proteins. Four groups of seven C57BL/6 mice each were inoculated intravenously with 1 x 105 MB16F0 melanoma cells, and subsequently treated with five daily injections of the recombinant proteins comprising human IL-2-derived beta super-binders fused to human Fc(LALA). Two additional groups were treated with similarly formatted proteins containing either human IL-2(K35E) or H9 superkine. Mice from a control group received PBS injections. All the animals were sacrificed at day 21, and their lungs were extracted for counting the number of metastatic nodules. Two animals, from the groups treated with PBS and IL-2(K35E)Fc(LALA) respectively, died before completion of the experiment and were excluded. Lines represent mean values and SD within each group.

Figure 10c. Figure 10. Anti-tumour effects of IL=2=derived fusion proteins. Seven groups of five BALB/c mice each were inoculated subcutaneously with 1 x 105 CT26 cells in the left flank, and subsequently treated with five daily injections of the set of recombinant proteins comprising human IL-2-derived beta super-binders, wild-type IL-2 (K35E) and H9 superkine fused to human Fc(LALA), and with PBS in the case of the control group. After tumours became palpable, their length and width were measured every other day in order to calculate tumor volume until one tumor side reached 18 mm. Tumor volumes are represented in c. Lines represent mean values and SD within each group