*2.4. C2\_2E12 Can Neutralize IL-33*/*ST2 Axis Driving Downstream Signaling Pathway in Human Cell Line*

To corroborate the neutralizing effects of C2\_2E12 in the IL-33/ST2 signaling axis in cells, we tested IL-33 induced MAPK and NF-κB pathways activation in human mast cells (HMC-1). HMC-1 cells expressed endogenous ST2 receptor and IL-1RAcP co-receptor, unlike HeLa cells (Figure 4A). The levels of phosphorylated extracellular signal-regulated kinase (ERK) and c-Jun N-terminal kinase (JNK) in the MAPK pathway were increased by the IL-33 C208S/C232S:ST2 complex while the level of inhibitor of NF-κB α subunit (IκBα) in the NF-κB pathway was reduced, supporting that IL-33 activates both MAPK and NF-κB pathways (Figure 4A). Subsequently, we treated the HMC-1 cells with IL-33 C208S/C232S alone or pre-incubated with C2\_2E12 and analyzed phosphorylation levels of ERK, JNK, and IκBα (Figure 4B,C). The relative phosphorylation level of IκBα is decreased by C2\_2E12 in a dose-dependent manner, indicating the suppression of the NF-κB signaling pathway. Relative phosphorylation levels of ERK and JNK were reduced by C2\_2E12 in a dose-dependent manner, implicating the suppression of the MAPK pathway. Analysis of the results indicated that C2\_2E12 can neutralize IL-33 and ST2 interaction by binding with IL-33 in a dose-dependent manner (Figure 4B,C). Taken together, our results

demonstrate that C2\_2E12 treatment can reduce IL33/ST2 complex formation by interfering with IL-33 and ST2 binding and thus act as a neutralizing antibody for the suppression of the IL-33/ST2 signaling axis in cells.

**Figure 4.** Intervention of IL-33/ST2 signaling axis by C2\_2E12 antibody. (**A**) IL-33-mediated activation of nuclear factor κB (NF-κB) signaling in human mast cells (HMC-1) cells, but not in HeLa cells where ST2 and IL-1RAcP were not expressed. HMC-1 and HeLa cells were stimulated with mock or human IL-33 (1 ng·mL<sup>−</sup>1, 8 min), and cell lysates were analyzed by immunoblotting with the indicated antibodies. (**B**) The inhibition of IL-33-induced NF-κB signaling by C2\_2E12 antibody in a dose-dependent manner. HMC-1 cells were treated with IL-33 alone or IL-33 pre-incubated with the increasing amounts of C2\_2E12 antibody. Cell lysates were analyzed by immunoblotting with the indicated antibodies. IL-33 (1 ng·mL<sup>−</sup>1) was pre-incubated with increasing amounts of the C2\_2E12 antibody (0.1, 0.5, and 2.0 ng·mL−<sup>1</sup> in lanes 3, 4, and 5, respectively) for 15 min and treated to HMC-1 cells for 8 min. (**C**) Quantification of protein levels shown in (**B**). Band intensities in immunoblots were quantified using ImageJ. Data are expressed as the mean ± SEM. The statistical significance of differences was analyzed by one-way analysis of variance (ANOVA) followed by Bonferroni's multiple comparison test (\*\*\* *p* < 0.001 compared to the indicated points; *n* = 3).

### **3. Discussion**

IL-33 has been associated with several chronic diseases such as asthma, atopic dermatitis, and inflammatory allergy, and recently, it was discovered that it plays important roles in regulatory immune responses. Neutralizing antibodies against IL-33 or ST2 have been developed to hinder IL-33 and ST2 binding. In this study, we discovered an scFv that specifically binds to IL-33 and subsequently interferes with IL-33 and ST2 complex formation. The IL-33 epitope region with C2\_2E12 overlaps the ST2 binding domain in IL-33, implicating that C2\_2E12 can act as a neutralizing antibody by competitive binding to IL-33. Pull-down assay and human cell line analysis verified that C2\_2E12 has neutralizing efficacy. IL-33 and ST2 interaction stimulates the activation of immune cells such as mast cells, T-helper type 2 cells, and dendritic cells, thereby causing allergic inflammatory responses. Therefore, future evaluation of C2\_2E12 and its refined clone(s) would desirably include the determination of efficacies in the keratinocyte/dendritic cells or epithelium/dendritic cell co-cultures. IL-33 is also known to stimulate type 2 innate lymphoid cells to release cytokines such as IL-5 and IL-13 [25]. Determination of the secreted IL-5 and IL-13 in response to the intervention of IL-33/ST-2 signaling axis by C2\_2E12 would require future studies. C2\_2E12 is not only a neutralizing antibody binding to IL-33, but also a monoclonal antibody discovered from human synthetic library of scFvs in vitro. The identification of epitope at the residue level, neutralizing the efficacy and human origin of C2\_2E12 renders it a suitable candidate for further engineering. Our preliminary comparative data reveal that C2\_2E12 shows an affinity only marginally inferior to that of a commercial antibody (data not shown). After further improving the affinity of C2\_2E12 by affinity maturation and going through an in vivo test, C2\_2E12 and its refined clone(s) could be used as a therapeutic antibody against IL-33 for treating allergic inflammatory diseases.

#### **4. Materials and Methods**

#### *4.1. Plasmid Constructs Cloning*

Genes encoding the mature form of IL-33 (residues 112–270), hereafter called simply "IL-33", and the ectodomain of ST2 (residues 19–321) were synthesized (Cosmo Genetech, Seoul, Korea) and cloned into BamHI/StuI sites of parallel GST-2 vector [26] and *Not*I/*Nco*I sites of the pSF vector, respectively. CH2 and CH3 domains of human the IgG Fc region were cloned into XhoI/BsgI sites of pSF-ST2 plasmid for ST2-Fc fusion protein expression. Plasmids encoding IL-33 mutants (deletion mutants, alanine-scanning mutants, and an oxidation-resistant mutant C208S/C232S) were prepared by following the protocol for QuikChange kit (Agilent, Santa Clara, CA, USA). Identities of all the constructs were verified by DNA sequencing.

#### *4.2. Expression and Purification of Recombinant Proteins*

The plasmid encoding GST-IL-33 was transformed into *E. coli* BL21 (DE3) cells. A single colony was inoculated into 10 mL Luria broth (LB) media containing 100 <sup>μ</sup>g·mL−<sup>1</sup> ampicillin and grown at 37 ◦C overnight. After 16–18 h, the pre-cultured cells were transferred to 500 mL LB media containing 100 <sup>μ</sup>g·mL−<sup>1</sup> of ampicillin, grown at 37 ◦C until OD600 0.6–1.0, induced with 0.6 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG), and further grown at 25 ◦C overnight with gentle shaking. Cells were harvested by centrifugation and re-suspended in lysis buffer (50 mM Tris-HCl pH = 7.5 and 150 mM NaCl). Cells were disrupted by ultrasonication, cleared by centrifugation, and the supernatant containing GST-IL-33 was transferred to Glutathione Sepharose 4B resin (GE Healthcare, Chicago, IL, USA) pre-equilibrated with the lysis buffer. After washing the resin with the lysis buffer, GST-IL-33 was eluted in GST elution buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, and 10 mM reduced glutathione). IL-33 was relieved from the GST fusion protein using a recombinant His6-tagged tobacco etch virus (TEV) protease during dialysis at 4 ◦C overnight. The resulting IL-33 was further purified on a Superdex 75 10/300 GL size-exclusion chromatography column (GE HealthCare, Chicago, IL, USA) pre-equilibrated with the lysis buffer. A ST2-Fc fusion protein, containing the ectodomain (residues

19–321) of ST2 and Fc from human IgG, was expressed in Expi293F cells (Thermo Fisher Scientific, Waltham, MA, USA) maintained in Expi293 expression medium (Thermo Fisher Scientific, Waltham, MA, USA). The day before transfection, cells were seeded to a final density of 2 <sup>×</sup> <sup>10</sup><sup>6</sup> viable cells ml−<sup>1</sup> in a 125 mL Erlenmeyer flask and grown at 37 ◦C for 24 h. After 24 h, cells were transfected with 30 μg of pSF-ST2-Fc plasmid DNA diluted in Opti-MEM™ I Medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 80 μL of ExpiFectamine™ 293 Reagent (Thermo Fisher Scientific, Waltham, MA, USA). Four days post-transfection, the cells were harvested by centrifugation, and its supernatant containing ST2-Fc was transferred to protein A agarose (Thermo Fisher Scientific, Waltham, MA, USA) pre-equilibrated by phosphate-buffered saline (PBS). The resin was washed using PBS, and ST2-Fc was eluted in Fc elution buffer (100 mM glycine pH = 3 with 1/10 volume of 1 M Tris-HCl pH = 8.0). The eluted ST2-Fc was dialyzed in the lysis buffer at 4 ◦C overnight.

#### *4.3. Biopanning Using Phage Display*

A synthetic human scFv library encoding His6- and HA-tagged scFv clones was used for biopanning [27]. Biopanning was performed as described previously with some modifications [28]. Library biopanning was performed in immuno tubes coated with the recombinant GST–IL-33 as an antigen in two conditions: inclusion of a negative selection with GST in every round (condition 1), and in the first round only (condition 2). In condition 1, five rounds of biopanning were performed in immuno tubes coated with GST protein for a negative selection and the recombinant GST-IL-33 at a gentle decrease in antigen concentrations (50, 10, 7.5, 5, and 2.5 <sup>μ</sup>g·mL<sup>−</sup>1). In condition 2, the negative selection using GST protein is performed only in the first round of biopanning and coated with GST-IL-33 in the other four rounds of biopanning using the same concentration with condition 1. To select scFv clones that specifically bind to IL-33, single colonies from the final round of a biopanning output plate were grown in a 96-well cell culture plate until OD600 reached 0.6–1.0 and induced with 1 mM IPTG grown overnight at 30 ◦C with shaking. The harvested cells in each 96-well were re-suspended in cold 1× TES buffer (50 mM Tris-HCl pH = 8.0, 1 mM ethylenediaminetetraacetic acid (EDTA) and 20% (*w*/*v*) sucrose) for 30 min on ice, and cold 0.2×TES buffer was added to the re-suspended cells for 1 h on ice. The recombinant GST-IL-33 protein at 10 <sup>μ</sup>g·mL−<sup>1</sup> in PBS was coated on a 96-well ELISA plate. ELISA assay for scFv screening with horseradish peroxidase (HRP) conjugated anti-HA secondary antibody (1:3000 dilution, Santa Cruz Biotechnology, Dallas, TX, USA) was performed with a final reading of signals recorded at OD450. The OD450 values with GST-IL-33 were divided by the OD450 value with GST, and the ratio of OD450 values was compared.

#### *4.4. Expression and Purification of scFvs in E. coli*

Cells were pre-cultured from the single colonies of scFv-expressing *E. coli* BL21 (DE3) at 37 ◦C overnight, transferred to 500 mL super broth (SB) media containing 100 <sup>μ</sup>g·mL−<sup>1</sup> ampicillin. Cells were grown at 37 ◦C until OD600 0.5–0.8 and induced with 1 mM IPTG at 30 ◦C with vigorous shaking. After 16–18 h, cells were harvested by centrifugation and re-suspended in cold 1× TES buffer for 30 min on ice, and cold 0.2× TES buffer was added to the re-suspended cells for 1 h on ice. The re-suspended cells supplemented with 5 mM MgCl2 to block EDTA were centrifuged, and their supernatants containing each scFvs were transferred to Ni-NTA agarose resin (Qiagen, Hilden, Germany). Each resin was washed by wash buffer A (PBS supplanted with 20 mM imidazole and 0.5 mM DTT) and scFvs were eluted by His-tag elution buffer A (PBS supplemented with 300 mM imidazole and 0.5 mM DTT). Size exclusion chromatography was performed on a Superdex 75 increase 10/300 GL column (GE HealthCare, Chicago, IL, USA) pre-equilibrated with PBS.

#### *4.5. Reformatting, Expression, and Purification of C2\_2E12 in Mammalian Cells*

The gene encoding C2\_2E12 in the pComb3X vector [27] was cloned into the pSF vector to express three different formats of C2\_2E12 in mammalian cells: scFv, antigen-binding fragment (Fab), and immunoglobulin G (IgG). The C2\_2E12 scFv was cloned into NotI/XhoI sites of the pSF vector for

expression in Expi293F cells (Thermo Fisher Scientific, Waltham, MA, USA) maintained in Expi293 expression medium (Thermo Fisher Scientific, Waltham, MA, USA). A variable heavy chain and variable light chain of C2\_2E12 were cloned into NotI/NcoI sites and HindIII/XhoI sites of the pSF vector, respectively, for the expression of Fab and IgG formats in Expi293F cells (Thermo Fisher Scientific, Waltham, MA, USA). Transfection and preparation steps for Fab and IgG were the same with those for scFv except for the amount and ratio of DNA plasmid used for transfection. The expression of scFv, Fab, and IgG formats of C2\_2E12 in mammalian cells was performed in the same way as for ST2-Fc. Supernatants containing C2\_2E12 scFv and Fab were transferred to Ni-NTA agarose resin (Qiagen, Hilden, Germany) pre-equilibrated with PBS. Each resin was washed by wash buffer B (PBS supplemented with 20 mM imidazole) and eluted by His-tag elution buffer B (PBS supplemented with 300 mM imidazole). Supernatant containing C2\_2E12 IgG was transferred to a protein A agarose (Thermo Fisher Scientific, Waltham, MA, USA) pre-equilibrated by PBS. The resin was washed using PBS, and the C2\_2E12 IgG was eluted in the Fc elution buffer. The antibodies were dialyzed at 4 ◦C overnight, concentrated, and loaded to a Superdex 75 10/300 GL size-exclusion chromatography column (GE HealthCare, Chicago, IL, USA) pre-equilibrated with PBS.

#### *4.6. Immunoblot*

Purified proteins were subjected to 12% SDS-PAGE and transferred onto polyvinylidene difluoride (PVDF) membranes (Merck Millipore, Burlington, MA, USA). Membranes were blocked with 5% (*w*/*v*) skim milk in Tris-buffered saline (pH = 7.5) containing 0.1% (*v*/*v*) tween-20 for 1 h at room temperature. Different primary antibodies were incubated at 4 ◦C overnight, and secondary antibodies were incubated at room temperature for 1 h. The bound antibody was detected by enhanced chemiluminescence (ECL) reaction with EZ-Western Lumi Pico kit (DoGen, Seoul, Korea).

### *4.7. Enzyme-Linked Immunosorbent Assay (ELISA)*

GST-IL-33 was coated on half the total area of a Costar® 96-well plate in the clear flat bottom polystyrene high bind microplate (Corning, Corning, NY, USA) and incubated at 4 ◦C for overnight. In the next day, the resulting culture was washed using PBST (PBS supplemented with 0.1% (*v*/*v*) tween-20) and blocked using blocking buffer (5% (*w*/*v*) skim milk in PBST). Serial dilutions of purified scFvs as the primary antibody with approximately 8 ng·mL−<sup>1</sup> to 0.8 mg·mL−<sup>1</sup> were added to the wells. The plate was incubated at ambient temperature for 1 h and washed using PBST. Subsequently, HRP-conjugated anti-HA antibody as the secondary antibody (1:3000 dilution, Santa Cruz Biotechnology) was added to the wells and incubated at ambient temperature for 1 h. The incubated plate was washed using PBST, and tetramethylbenzidine (TMB) substrate solution (GenDEPOT, Katy, TX, USA) was added for color development. After incubation for 10 min, 1 M H2SO4 was added to the plate to stop the color development reaction. The final signal readings were recorded at 450 nm and plotted using Prism 5 (GraphPad, San Diego, CA, USA).

#### *4.8. Biolayer Interferometry (BLI)*

Binding kinetics was measured by BLI experiments using a BLItz system (ForteBio, Fremont, CA, USA). *E. coli* cell lysate containing His6-tagged antibodies was prepared in 0.5× TES buffer. GST-IL-33 was prepared in BLI buffer (PBS supplemented with 20 mM imidazole, 0.05% (*v*/*v*) Triton X-100 and 0.1 mg·mL−<sup>1</sup> BSA) to reduce the nonspecific binding signal. The BLI buffer was also used as the kinetics buffer. The cell lysate was immobilized to Ni-NTA biosensors (ForteBio, Fremont, CA, USA) and washed using the kinetics buffer. The sensors were subsequently reacted with various concentrations (2, 1, 0.5, and 0.25 μM) of GST-IL33 (association step) and washed using the kinetics buffer (dissociation step). These assays were performed twice each. All real-time recorded sensograms were analyzed by the 'global fitting' method in BLItz Pro 1.2 (ForteBio, Fremont, CA, USA) to calculate *kon* (association

rate constant) and *ko*ff (dissociation rate constant) values. The *K*<sup>d</sup> (dissociation constant) value of each antibody was calculated using the following equation:

$$K\_{\rm df} = k\_{off} / k\_{on}.\tag{1}$$

*r*<sup>2</sup> analysis, an indication of goodness of graph curve fitting, was performed using BLItz Pro 1.2, and the *r*<sup>2</sup> values of all experiments were above 0.98. The graphs of raw sensograms were prepared by Prism 5 (GraphPad, San Diego, CA, USA).

### *4.9. Structural Modeling*

A homology structural model for C2\_2E12 that was generated using the SWISS-MODEL server [24] and the crystal structure of IL-33:ST2 complex (PDB ID: 4KC3) were used as templates for the docking of C2\_2E12 to IL-33. Protein–protein docking modeling was performed using the HADDOCK server [29]. To perform the HADDOCK modeling, a restraint was applied such that the two key epitope residues of IL-33, L150 and K151, must interact with the complementary determination region of C2\_2E12. The Z-score of clustering and other modeling parameters are listed in Table S2. Structural analysis of the interface between C2\_2E12 and IL-33 was performed using PyMOL 1.8 (Schrödinger, New York, NY, USA).
