Cloning and functional characterization of novel human neutralizing anti-interferon-alpha and anti-interferon-beta antibodies

Abstract

Type I IFNs play a pivotal role in immune response modulation, yet dysregulation is implicated in various disorders. Therefore, it is crucial to develop tools that facilitate the understanding of their mechanism of action and enable the development of more effective anti-IFN therapeutic strategies. In this study, we isolated, cloned, and characterized anti-IFN-α and anti-IFN-β Abs from peripheral blood mononuclear cells of individuals treated with IFN-α or IFN-β, harboring confirmed neutralizing Abs. Clones AH07856 and AH07857 were identified as neutralizing anti-IFN-α-specific with inhibition against IFN-α2a, -α2b, and -αK subtypes. Clones AH07859 and AH07866 were identified as neutralizing anti-IFN-β1a-specific signaling, and able to block Lipopolysaccharide or S100 calcium binding protein A14-induced IFN-β signaling effects. Cloned Abs bind rhesus but not murine IFNs. The specificity of inhibition between IFN-α and IFN-β suggests potential for diverse research and clinical applications.

Introduction

IFNs are a family of widely expressed and related cytokines that have potent antiviral, antiproliferative, antitumor and immunomodulatory activities (1–4). They act by inducing innate and promoting adaptive immune responses, and thus serve as the first line of defense against viral, bacterial infections, and malignant cells (5, 6). The IFN family consists of distinct proteins that fall into four discrete classes (type I, II, III, and IV) differentiated by their receptor complexes (7–11). Three of those classes (type I, II, and III) are found in human.

The type I IFN family includes 13–14 IFN-α subtypes which are involved in the innate immune response to infection, tumor development, and other inflammatory stimuli (12, 13). All the members of the type I IFN family bind to a common heterodimeric receptor (IFNα/β receptor, IFNAR) consisting of the low affinity IFNAR1 and the high affinity IFNAR2 subunits (6, 14, 15). Binding of type I IFNs to IFNAR2 results in the subsequent recruitment of IFNAR1, which leads to the formation of a high-affinity IFNAR. Although all type I IFN subtypes start signaling through the same type I IFNAR, the stability of binding between various type I IFN subtypes with IFNAR is different. IFN-α1 binds with the lowest affinity to both chains of IFNAR, while IFN-α2 binds to IFNAR1 with lesser (micromolar) affinity than to IFNAR2 (nanomolar) (16). On the other hand, IFN-β has a greater affinity for IFNAR1 and can also bind to IFNAR1 independently of IFNAR2 leading to transduction of specific, unconventional intracellular signals and contributing to toxicity in vivo (6, 17).

Binding of type I IFNs to IFNAR results in the subsequent activation mainly of the JAK-STAT pathway, and the transcription of IFN-stimulated genes [ISG, e.g. MX Dynamin Like GTPase 1 (MX1), Ubiquitin specific peptidase 18 (USP18)], which in turn allows cells to respond to IFN stimulation (4, 18–22). The activities of type I IFNs can be grouped to robust, and tunable. Robust activities (e.g. activities associated with anti-viral responsiveness) are common to all cells, require only minute amounts of all type I IFNs, and are independent of their binding affinity or the numbers of IFNAR on cell surface. On the other hand, tunable activities (e.g. activities observed after activation of robust activities and associated with the immunomodulatory and anti-proliferative effects of type I IFNs) are cell type-specific, require 1000-fold higher concentration of type I IFNs and higher number of IFNAR, are observed after longer times of treatment with type I IFNs, and are most strongly activated by the high affinity binding of IFN-β (or IFN-α2 variants engineered for tight binding) to IFNAR (15).

Type I IFNs have been used widely in the clinic primarily in the treatment of viral infections and of autoimmune conditions such as multiple sclerosis (MS). Regarding viral infections, IFN-α2a/b has been used to treat hepatitis virus B and C (23–25). In addition, in the context of HIV infection treatment with IFN-α2a/b was high in the pre-antiretroviral therapy (ART) era (26, 27) but was abandoned after wide usage of ART. Recent reports though, suggest a potential beneficial role of IFN-α2a/b in activating NK anti-viral activity in HIV cure-directed strategies (28–32). While antiviral strategies have pursued robust activities of type I IFNs via IFN-α2a/b, therapy for MS has focused on tunable activities of decreasing T cell activation as exemplified by usage of IFN-β (33, 34). Therapy with type I IFNs can result in the development of natural anti-IFNs Abs such as natural anti-IFN-β Abs in patients treated with IFN-β for MS that block IFN-β but do not block IFN-α (35–40), as well as development of natural anti-IFN-α Abs in patients treated with IFN-α (41–48).

Despite the central role of type I IFNs in the development of immune response and their great clinical importance, aberrant activation of the type-I IFN response has been shown to result in a large spectrum of disorders called interferonopathies, while in the context of HIV and SIV infection type I IFNs have been described to have beneficial and detrimental roles (5, 26, 49–56). Limited information has defined whether these differential outcomes are linked to specific type I IFN subtypes. However, in both viral and cancer a potential detrimental role has been linked to IFN-α or IFN-β expression (53, 57–60) supporting development of novel strategies for the selective inhibition of specific type I IFNs signaling. While human studies targeting in the same study specific type I IFN subtypes IFN-α versus IFN-β have yet to be performed, a larger number of studies has been conducted testing the effectiveness of inhibitors of all type-I IFN signal transduction, of mAbs against IFN-α, and of IFNAR antagonists (61–65). Here, we identify, clone and characterize novel natural neutralizing human anti-IFN-α and anti-IFN-β Ab clones derived from a single pooled PBMC cDNA from persons treated with IFN-α or IFN-β after having confirmed the respective presence of plasma neutralizing anti-IFN-specific Abs.

Materials and Methods

Participants

Cryopreserved plasma from 11 participants with MS, and from 18 persons living with HIV was screened for presence of anti-IFN-α and anti-IFN-β Abs respectively (Supplemental Table I). The persons living with HIV were on suppressive ART (<50 HIV-1 copies/ml for ≥1 year, and ≥450 CD4⁺ T cells/mm³ at study entry) and received IFN-α as immunotherapy for 7.25 months under a research protocol NCT00594880 study (28), while the patients with MS were treated with IFN-β for a maximum of 6 months. A total of 220×10⁶ and 346×10⁶ cryopreserved PBMCs from two selected patients (MS-001 and Peg-008 respectively) with activity against IFN-β or IFN-α respectively were used for the cloning of anti-IFN-β and anti-IFN-α Abs as described below. The study protocol and informed consent procedures were approved by the Institutional Review Boards of the authors’ institutions.

Cell lines

Human embryonic kidney (HEK) IFNAR1/IFNAR2 cell line was from Eurofins DiscoverX (catalog no. 93–1097C1). Expi293F cells were from Thermo Fisher Scientific Inc (catalog no. A14527). THP-1 cells were from Invivogen (catalog no. thp-null; San Diego, CA). HEK IFNAR1/IFNAR2 cells were incubated at 37°C in DMEM (catalog no. 10–013-CM; Corning, New York, NY) containing 10% FBS (catalog no. 35–010-CV; Corning), 0.8 mg/ml G418 (catalog no. 10131035; Thermo Fisher Scientific Inc, Philadelphia, PA), and 0.2 mg/ml Hygromycin (catalog no. 10687010; Thermo Fisher Scientific Inc). Expi293F cells were grown in serum-free Expi293^™ expression medium (catalog no. A39251; Thermo Fisher Scientific Inc) and maintained in Erlenmeyer Flasks (catalog no. 4117–0500; Thermo Fisher Scientific Inc) at 37°C with 8% CO₂ on an orbital shaker. THP1 cells were cultured at 37°C with 8% CO₂ in RPMI 1640 (catalog no. 0040CM; Cellgro Technologies LLC, Lincoln, NE), containing 5% FBS (Catalog no. 35–010-CV; Corning), 1% Penicillin and 1% Streptomycin (catalog no. 30002CI; Cellgro Technologies LLC).

Receptor dimerization assay

The presence of Abs against IFN-α or IFN-β was assessed using the IFN receptor dimerization assay. In this assay, IFN receptor dimerization is detected upon IFN binding in an engineered HEK cell line (HEK IFNAR1/IFNAR2) using enzyme fragment complementation technology (catalog no. 93–1097C1; Eurofins DiscoverX, Fremont, CA). This cell line co-expresses IFNAR1 and IFNAR2, each conjugated to a fragment of β-galactosidase. Induction of receptor dimerization by IFN-α or IFN-β brings the enzyme fragments together to form an active enzyme, which hydrolyzes its substrate and generates a detectable chemiluminescence signal (Eurofins DiscoverX). The presence of anti-IFN-α or anti-IFN-β Abs, which bind to IFN-α or IFN-β, respectively, prevents receptor dimerization and the subsequent generation of chemiluminescence signal. HEK IFNAR1/IFNAR2 cells were maintained in DMEM (catalog no. 10–013-CM; Corning, New York, NY) containing 10% FBS (catalog no. 35–010-CV; Corning), 0.8 mg/ml G418 (catalog no. 10131035; Thermo Fisher Scientific Inc, Philadelphia, PA), and 0.2 mg/ml Hygromycin (catalog no. 10687010; Thermo Fisher Scientific Inc). Briefly, cells were seeded in half-area white 96-well microplates (20000 cells/well) and incubated at 37°C overnight. Meanwhile, plasma samples or cloned anti-IFN-α or anti-IFN-β Abs were pre-incubated with human IFN-a2a (IFN-α, catalog no. 11100–1; PBL Assay Science, Piscataway, NJ) or human IFN-β1a (IFN-β, catalog no. 11415–1; PBL Assay Science) at 4°C overnight. For IFN-β, plasma was pre-incubated overnight with IFN-β at 4°C at 3X of the final concentrations, while for IFN-α, plasma was first heat inactivated at 56°C for 1 h and then pre-incubated overnight with IFN-α at 4°C, at 3X of the final concentrations. Final concentrations in the wells with HEK IFNAR1/IFNAR2 cells were as shown: 11% plasma, or 67% clone supernatant, 2.4 nM IFN-α, and 1.2 nM IFN-β. After the overnight incubation, plasma pre-incubated with IFN-α or IFN-β was added to the cells and allowed to incubate at 37°C for 24 hs. Added controls included plasma samples from two healthy individuals, F623-C32 and M486-B55 as negative controls, as well as expressed/purified humanized IgG CTI-AF1 (anti-IFN-β Ab, 10 μg/ml; patent US20170313769A1) or BMS-13H5 (neutralizing anti-IFN-α-specific Ab, 1.1 μg/ml or 10 μg/ml; patent W02005059106A2) (66, 67) as positive controls, respectively. Luminescence signal was measured using the Envision microplate reader (PerkinElmer, Waltham, MA). The identification of samples with anti-IFN-α or anti-IFN-β Abs was based on the inhibition of IFN receptor dimerization despite the presence of IFN-α or IFN-β. Samples were run repeatedly in duplicates.

Ab depletion

For Ab depletion, plasma sample at 10% in complete media was incubated with Protein A/G sepharose beads (catalog no. ab193262; Abcam, Boston, MA) for 1 h with rotation at 4°C, for 5 consecutive times. For the receptor dimerization assay, the Ab depleted plasma samples were pre-incubated with IFN-β or IFN-α at 4°C overnight. Subsequently, HEK IFNAR1/IFNAR2 cell media was replaced by the Ab depleted plasma pre-incubated with IFN-β or IFN-α and incubated at 37°C for 24 hs (final concentrations: 10% plasma, 1 nM IFN-β or 2.4 nM IFN-α). Luminescence signal was measured in a microplate reader. Detection of depletion of anti-IFN-β and anti-IFN-α Abs in the plasma sample by Western blotting was also performed. To detect Ab in the plasma sample, rabbit anti-human IgG (H+L) [catalog no. 309–005-003; Jackson Immuno-Research (JIR), Baltimore, PA] at 1:5000 was used as the primary Ab, and 800 CW goat anti-rabbit IgG (catalog no. 926–32211; LI-COR, Lincoln, NE) at 1:5000 as the secondary Ab. Albumin detection in the serum samples after consecutive rounds of Ab depletion was used to ensure the depletion was Ab specific. Following membrane stripping goat anti-human/mouse serum albumin (catalog no. AF3329; R&D Systems Inc, Minneapolis, MN) at 1:2500 was used as the primary Ab, and 800CW donkey anti-goat IgG (catalog no. 926–32214; LI-COR) at 1:5000 as the secondary Ab.

IFN-mAb development and recombinant mAb generation

Total RNA was isolated from PBMC samples using RNeasy kit (catalog no. 74004; Qiagen, Germantown, MD) and pooled. After reverse transcription from pooled RNA using a reverse transcription-PCR kit (catalog no. 12596100; Thermo Fisher Scientific Inc), a master cDNA preparation was synthesized as a template to develop human Fab phagemid library. The VH and Fab-VL (Vk or Vl) genes were amplified by PCR followed by nested PCR using mixtures of primers specific for IgG and Igk and Igl. Amplified products from first-round and nested primer sets were then sequenced or IgG-VH and IgG-VL.

Recombinant Plasmid Preparation and Ab expression

Recombinant Abs were cloned, as previously described (68). To generate recombinant Abs, H chain and L chain fragments were cloned into pCDNA 3.4 Abs expression vectors (catalog no. A14697; Thermo Fisher Scientific Inc) in frame with either human IgG (IgGk) constant domain.

For expression in mammalian cell lines, Ab constructs were first cloned into CMV expression vectors which relies on the widely utilized human CMV intermediate-early promoter/enhancer to drive overexpression of cloned inserts. Gene constructs encoding full length IgG, and LALA mutation are designed in silico and obtained from Genscript Inc (Piscataway, NJ) offering gene synthesis services. Transfection grade plasmids were maxi-prepared, for Expi293F cell expression (catalog no. A14527; Thermo Fisher Scientific Inc). Expi293F cells (100 ml) were grown in serum-free Expi293^™ expression medium (catalog no. A39251; Thermo Fisher Scientific Inc). Following large-scale preparation of plasmid DNA, we used a suspension adapted HEK293 cell line (Expi293F) for high levels of transient expression. Optimization of heavy-to- light chain transfection ratios (IgG and Fab formats) maximized recombinant Ab yields. Cells were maintained in Erlenmeyer Flasks at 37°C with 8% CO₂ on an orbital shaker. One day before transfection, the cells were seeded at an appropriate density. On the day of transfection, DNA and transfection reagent were mixed at an optimal ratio and then added into the flask with cells ready for transfection. Recombinant plasmids encoding target Ab were transiently co-transfected into Expi293F cell cultures. Cell culture supernatants collected on day 6 were used for Ab purification.

ELISA and Western blot for hybridomas screening

ELISA was carried out using the 96-well MaxiSorp plates (catalog no. 442404; Nunc Thermo Fisher Scientific Inc) coated with 1 μg/ml (volume 100 μl/well) of human recombinant IFN specific proteins (R&D Systems) in PBS, pH 7.4 and incubated at 4°C overnight. Primary Abs were added at 1:50 dilution. Following incubation, plates were washed with PBS-T [1xPBS (catalog no. 21–030-CM; Cellgro Technologies LLC, Lincoln, NE) with 0.05% Tween 20 (catalog no. P9416; Sigma-Aldrich, St Louis, MO)] and blocked using PBS (catalog no. 21–030-CM; Cellgro Technologies LLC) containing 10% FBS (catalog no. 35–010-CV; Corning) for 1 h at room temperature (RT). Subsequently, the plates were washed with PBS-T and incubated with hybridoma, serially diluted in PBS (catalog no. 21–030-CM; Cellgro Technologies LLC) with 1% FBS (Corning) and 0.1% Tween 20 (catalog no. P9416; Sigma-Aldrich) for 30 min on a shaker and 90 mins at RT. After another wash, the plates were treated with goat-anti-mouse IgG1 (catalog no. A90–105A; Bethyl Laboratories, Montgomery, TX) at a dilution of 1:10000 at RT for 1 h. Post final wash, the plates were developed with SigmaFast OPD substrate (catalog no. P9187–5SET; Sigma-Aldrich) for 5–10 min in the dark, and the reaction was stopped using 1 N H₂SO₄. (catalog no. SX1244–6; Sigma-Aldrich). The plates were read using a Synergy2 plate reader (BioTek Instruments, Winooski, VT) at an OD of 450nm (68).

For the avidity test, the in-house avidity assay was standardized using a commercial ELISA kit (catalog no. BMS2091; Thermo Fisher Scientific Inc) for detecting specific IgG Abs modified to incorporate an elution step with urea to remove low-avidity Abs from target antigen. For the assay, 100 μl of each serum diluted were added to wells of polystyrene plates coated with human IFN-α2a (catalog no. 11101–2; R&D Systems Inc) and IFN-β antigens (catalog no. 8499-IF; R&D Systems Inc). All serum samples were run twice in duplicate, as described before (68).

Western analysis was used to evaluate detection of IFNs and to analyze the light/heavy chain composition of cloned Ab preparations. For IFN detection, human recombinant IFN-α2a (1 μg/ml; catalog no. 11101–2; R&D Systems Inc) and IFN-β (1 μg/ml; catalog no. 8499-IF; R&D Systems Inc) proteins were reduced using NuPAGE Sample Reducing Agent (10x) (catalog no. NP0004; Thermo Fisher Scientific Inc) and heating at 70°C for 10 min, then loaded onto sample lanes with Odyssey Protein Molecule Weight (catalog no. 928–40000; LI-COR) serving as a standard marker. For cloned Ab composition evaluation, the mAbs following expression were purified and subjected to SDS-PAGE analysis using a NuPAGE 4–12% Bis-Tris gel (catalog no. NP04125BOX; Thermo Scientific Inc) under reducing (lane 1) and non-reducing (lane 2) conditions. For Western analysis, an added positive human IgG Kappa (catalog no. 15154; Sigma-Aldrich) loaded control lane was included (lane P). The gel electrophoresis was carried out using sodium dodecyl sulfate-12% polyacrylamide gel for 50 min at 200 V. Following electrophoresis, samples were transferred onto polyvinylidene fluoride membranes via an iBlot-2 system (catalog no. IB21001; Thermo Fisher Scientific Inc) and blocked using Odyssey Blocking Buffer (OBB) (catalog no. 927–60003; LI-COR) for 1–2 hs on a rocker. For IFN detection gels, membranes were treated with cloned Ab supernatant (1:200) in OBB (catalog no. 927–60003; LI-COR) containing 0.1% Tween 20 (catalog no. P9416; Sigma-Aldrich) at 4°C overnight. Following incubation, the membranes were washed four times at 5 min intervals with PBS-T. Subsequently, washed membranes were treated with IRDye 800 CW goat-anti-human secondary Ab (catalog no. 926–32210; LI-COR) in OBB (catalog no. 927–60003; LI-COR) containing 0.1% Tween 20 (catalog no. P9416; Sigma-Aldrich) and 0.01% SDS (catalog no. 151–21-3; Sigma-Aldrich) at a dilution of 1:10000 and incubated at RT for 60 min in the dark on the rocker. For Ab clone composition evaluations, primary Ab goat anti-human IgG-HRP (catalog no. A00166; GenScript Inc), and goat anti-human Kappa-HRP (catalog no. 2060–05; SouthernBiotech, Birmingham, AL) were used. Following incubation, the membranes were rewashed four times and scanned using Odyssey CLX Imager (LI-COR).

Assessment of the binding of the cloned anti-IFN-α or anti-IFN-β Abs on IFN-α or IFN-β by surface plasmon resonance

IFN Abs affinity was assessed by surface plasmon resonance (SPR) using a Biacore T200 (Cytiva Life Sciences, Marlborough, MA). Approximately, 300 response units (RU) of IFN-α [human (catalog no. 13833-HNAY), rhesus (catalog no. 90105-CNAY), or mouse (catalog no. 50525-MNAY) untagged, Sino Biological, Chesterbrook, PA] or 500 RU of IFN-β [human (catalog no. 10704-H02H), rhesus, (catalog no. 90104-C05H) or mouse (catalog no. 50709-M02H), C-terminal Fc tagged, Sino Biological] was immobilized on carboxymethyldextran sensor chip (catalog no. SCBS-CMD200M; Xantec Bioanalytics, Germany) using amine coupling. Briefly, the sensor chip was first washed with 0.1 M sodium borate, pH 9.0 (catalog no. SX0355; Sigma-Aldrich), 1 M sodium chloride (catalog no. 59222C; Sigma-Aldrich) for 3 min at 10 ml/min, followed by activation with 50 mM 3-Dimethylamino-propyl-ethyl-carbodiimide Hydrochloride (catalog no. 03449)/100 mM N-Hydroxysuccinimide (catalog no. 130672) (Sigma-Aldrich) for 12 min at 10 ml/min. Either IFN-α or IFN-β in 10 mM sodium acetate, pH 5.0 (catalog no. S2889; Sigma-Aldrich) was then flowed over the chip until the desired immobilization level was achieved. After a 30 min delay, the remaining activated sites were blocked with 1 M ethanolamine hydrochloride, pH 8.5 (catalog no. E6133; Sigma-Aldrich) for 5 min at 10 ml/min. The running buffer was then switched from distilled water to 10 mM HEPES, pH 7.4 (catalog no. 54457; Sigma-Aldrich), 150 mM sodium chloride (catalog no. 59222C; Sigma-Aldrich), 0.1% Tween20 (catalog no. P9416; Sigma-Aldrich). Antibodies were serially diluted 1:3.16 for 6 concentrations starting at 320 nM for IFN-α Abs, and 1000 nM for IFN-β Abs. The association and dissociation times were 180/420 sec for IFN-α Abs, and 360/600 sec for IFN-β Abs at 25 ml/min. After the respective dissociation time, the Ab was dissociated by a 60 sec injection of 20 mM glycine, pH 2.0 (catalog no. 50046; Sigma-Aldrich) at 30 ml/min. Kinetic parameters were obtained from global nonlinear regression fits of the data using Biacore evaluation software (Cytiva Life Sciences).

Assessment of the effect of the cloned anti-IFN-α or anti-IFN-β Abs on the IFN-α or IFN-β-mediated STAT-1 phosphorylation in primary cells by flow cytometry

Two million PBMC (10×10⁶ cells/ml) were added into 15 ml conical tubes and stained with surface mouse Abs [CD3-PE-Cyanine7 (PECy7); catalog no. BD557851, CD4-V450; catalog no. BD560345] or corresponding surface mouse isotypes (IgG1k-PECy7; catalog no. BD557872, IgG1k-V450; catalog no. BD560373) at 4°C for 30 min. Cells were then washed with RT 1XPBS (catalog no. 21–030-CM; Cellgro Technologies LLC) at 1500 rpm for 5 min and re-suspended in warm 1 ml 1XPBS (catalog no. 21–030-CM; Cellgro Technologies LLC). Subsequently, the cloned anti-IFN-α or anti-IFN-β Abs were incubated with human IFN-α (catalog no. 11100–1; PBL Assay Science) or human IFN-β (catalog no. 11415–1; PBL Assay Science) at RT for 15 min and then added to the cells at 37°C for 10 min. Cells were then fixed with 16% light sensitive paraformaldehyde (catalog no. RT15710; Electron Microscopy Sciences Inc, Hatfield, PA) at a final concentration of 5% paraformaldehyde at 37°C for 10 min. Subsequently, 10ml of 1XPBS were added and cells were centrifuged at 2200 rpm for 10 min. Supernatant was removed, the cell pellet was loosen up, and incubated with 500 μl – 1 ml of 1X PhosFlow Buffer (special order from BD Biosciences, San Diego, CA, USA) at RT for 30 min. Then, 3 ml FACS wash [1xPBS (catalog no. 21–030-CM ; Cellgro Technologies LLC) supplemented with 0.1% BSA (catalog no. A1470–100G; Sigma-Aldrich) and 0.02% NaN₃ (catalog no. 26628–22-8; Thermo Fisher Scientific Inc), supplemented with 5% human serum (catalog no. H4522; Sigma-Aldrich) and 5% mouse serum (catalog no. M5905; Sigma-Aldrich)] was added, cells were centrifuged at 2200 rpm for 10 min, the FACS wash was removed, intracellular mouse Stat1-PE (catalog no. BD612564) or corresponding intracellular isotype (IgG2ak-PE, catalog no. BD558595) was added in corresponding tubes and incubated in the dark at RT for 1 h. Then, 3 ml FACS wash were added, cells were centrifuged at 2200 rpm for 10 min, FACS wash was removed, cells were transferred into Facs tubes and run on LSR II Flow cytometer (BD Biosciences). A minimum of 100000 events were collected in the live cell gate (FSC/SSC). Data were analyzed using FloJo software (Version 8.8.4, Tree Star, Ashland, OR). The gating strategy for analysis of STAT-1 expression in CD3⁺CD4⁺ cells is shown in Supplemental Fig. 1. All Abs were from BD Biosciences.

Assessment of the effect of the cloned anti-IFN-α or anti-IFN-β Abs on the IFN-α or IFN-β-mediated STAT-1 phosphorylation, IFN-induced GTP-binding protein (MX-1) expression, and ubiquitin specific peptidase 18 (USP18) expression in THP-1 cells by Western blot

In order to assess the effect of cloned anti-IFN-α or anti-IFN-β Abs on IFN-α or IFN-β-mediated protein expression, THP-1 cells (Invivogen) were seeded in 6-well plates (1.5×10⁶ cells/well) with or without phorbol 12-myristate 13-acetate (PMA, 100 ng/ml, 2 ml/well, catalog no. 16561–29-8; Sigma-Aldrich) and incubated at 37°C for 48 hs. At the end of the 48 hs incubation, cloned anti-IFN-α or anti-IFN-β Abs were pre-incubated with human IFN-α2a [different subtypes of human IFN-α, human IFN-α sampler set, catalog no. 110021; PBL Assay Science), human IFN-β1a (IFN-β, catalog no. 11415–1; PBL Assay Science), LPS (catalog no. L4516; Sigma-Aldrich), or the S100 calcium binding protein A14 (S100A14, catalog no. S14-H5121; ACROBiosystems, Newark, DE) at RT for 30 min at RT, and then were added to the cells for an additional incubation at 37°C for 30 mins (for assessment of STAT-1 phosphorylation) or 24 hs (for assessment of MX1 and USP18 expression). At the end of the incubation the RPMI culture media [RPMI 1640 (catalog no. 10040CM; Cellgro Technologies LLC), 10% human serum (catalog no. H4522; Sigma-Aldrich), 100 U/ml Penicillin/100 μg/ml Streptomycin (catalog no. 30002CI; Cellgro Technologies LLC), 55 μM 2-Mercaptoethanol (catalog no. 21985–023; Gibco-Thermo Fisher Sicientific)] was replaced with fresh RPMI culture media and cells were further incubated at 37°C for 8 hs. Cells were subsequently washed with ice cold 1XPBS (catalog no. 21–030-CM; Cellgro Technologies LLC) twice and then lysed with 150 μl 1X cell lysis buffer (catalog no. 9803; Cell Signaling Technology, Danvers, MA), on ice for 10 min followed by removal of the lysed cells, sonication for 10 sec and centrifugation at 13000 g in a benchtop centrifuge at 4°C for 15 min. The protein concentration was determined in the samples using Pierce Bicinchoninic acid (BCA) Protein Kit (catalog no. 23227; Thermo Fisher Scientific Inc) and 20 μg of protein from each sample were used for Western blot.

Briefly, for immunoblotting, cells were lysed with 1x cell lysis buffer (catalog no. 9803; Cell Signaling Technology) containing protease inhibitor (catalog no. 11836170001; Roche, Basel, Switzerland) and phosphatase inhibitor (catalog no. 04906837001; Roche). The lysates were incubated on ice for 10 min followed by removal of the lysed cells and sonication for 10 sec, and then centrifugation at 13000 g in a benchtop centrifuge at 4°C for 15 min. The protein concentration was determined in the samples using the Pierce BCA Protein Kit (catalog no. 23227; Thermo Fisher Scientific Inc). The solid protein was denatured at 95°C for 5 min in NuPAGE LDS Sample Buffer (4X) (catalog no. NP0007; Thermo Scientific Inc) containing 50 mM DTT (catalog no. D9163; Sigma-Aldrich). Samples were loaded to 4–12% NuPAGE Bis-Tris Gels (catalog no. NP04125BOX; Thermo Scientific Inc) and run at 150 V for 1 h with NuPAGE MOPS SDS Running Buffer 20X (catalog no. NP0001; Thermo Scientific Inc). Following electrophoresis, the samples were transferred onto polyvinylidene fluoride membranes via an iBlot-2 system (catalog no. IB21001; Thermo Fisher Scientific Inc) and blocked using LI-COR Blocking Buffer (OBB, catalog no. 927–60003; LI-COR) for 1–2 hs on a rocker. The primary Abs were treated in OBB containing 0.2% Tween-20 overnight at 4°C on the rocker. The following primary Abs, used at 1:1000 dilution, were purchased from Cell Signaling Technology: rabbit pSTAT1 (catalog no. 7649S), rabbit STAT1 (catalog no. 9172S), rabbit MX1 (catalog no. 37849S), rabbit USP18 (catalog no. 4813S) and mouse β-actin (catalog no. 3700S). The secondary Abs, IRDye 800 CW (goat anti-rabbit IgG, catalog no. 926–32211), and IRDye 680 LT (donkey anti-mouse IgG, catalog no. 926–68022), were used at 1:5000 dilution, were purchased from LI-COR). The Western blot images were recorded using the Odyssey CLX imager (LI-COR) and the Western blot quantification was done using Image Studio (version 5.2).

Quantification and statistical analysis

Data were summarized as means and SD using JMP Pro15.2.0 (SAS Institute, Cary, NC). In the dimerization assays, means and SD of percent (%) inhibition was used, which was calculated by considering the luminescence value of the control sample without IFN (“no IFN”) as 100 % inhibition and the luminescence value of the control sample with IFN (“IFN only”) as 0% inhibition and by using the following formula: 100 − [(sample − “no IFN”) / (“IFN only” − “no IFN”) * 100]. In Western data means and SD of normalized data from 3 experiments were calculated. Statistical differences between different concentrations of each clone for multiple Western blots experiments in Fig. 5A were assessed by paired t tests, and two-sided p values <0.05 were considered statistically significant. Western data for each experiment were analyzed using Imager Studio (version 5.2) and first normalized to β-actin equal control for each sample, and then normalized to either no-stimulated control, or IFN-α or IFN-β alone without Ab control. The no-stimulation control was 100%, while the stimulation over the no-stimulation control was above 100%.

FIGURE 5.

Inhibition of IFN-mediated expression of pSTAT-1, MX1 and USP18 in THP-1 cells by neutralizing anti-IFN-specific Abs. Shown are Western blots analysis for the expression of pSTAT-1, STAT-1, MX1, USP18 and β-actin in THP-1 cells following stimulation for 30 mins for pSTAT-1 and STAT-1, and for 24 hs for MX1 and USP18 after IFN-α subtypes or IFN-β stimulation. (A) Bar graphs showing mean and standard error of three Western blots experiments measuring MX1 expression relative to β-actin levels in PMA-differentiated THP-1 cells following stimulation for 24 hs with 0.01 nM of IFN-α2a or IFN-β1a in the presence or absence of clones AH07856, AH07857, AH07859 and AH07866 at 30, 10, 3, and 0.3 μg/ml. Statistical significant differences between the different concentrations as compared to 0 μg/ml control for each clone for the 3 experiments are shown on the top of the graphs wth * if <0.05, ** if <0.001, and *** if <0.0001. (B) Bar graphs showing the results of the Western blots of one experiment measuring pSTAT-1/STAT-1, MX1 and USP18 expression relative to β-actin levels in PMA-differentiated THP-1 cells following stimulation for 30 mins for STAT-1, and 24 hs for MX1 and USP18 with 0.01 nM of IFN-α subtypes or 0.005 nM IFN-β1a in the presence or absence of clones AH07856, and AH07857 at 30, 10, 3, and 0.3 μg/ml. The Western blots for this experiment are shown in Supplemental Fig. 7. (C) Western blots showing expression of pSTAT-1, STAT-1, MX1, USP18 and β-actin in PMA-differentiated THP-1 cells following stimulation for 30 mins for pSTAT-1 and STAT-1, and for 24 hs for MX1 and USP18 after exposure to 0.005 nM of IFN-β1a in the presence or absence of clones AH07859 and AH07866 used alone or in combination at 30, 10, 3, and 0.3 μg/ml. (D) Bar graphs showing pSTAT-1/STAT-1, MX1 and USP18 expression relative to β-actin levels in PMA-differentiated THP-1 for the Western blots shown in panel (C). Western products evaluated were pSTAT-1/STAT-1: 91, 84 kDa, MX1: 76 kDa, USP18: 39, 34 kDa, and β-actin: 42 kDa. In all panels, IFN-α, IFN-α subtypes and IFN-α-specific clones are indicated with orange fonts, while IFN-β, IFN-β1a and IFN-β-specific clones are indicated with purple fonts.

Results

Identification of neutralizing Abs specific against IFN-β or IFN-α

We used the IFNR dimerization assay in the presence of IFN-β or IFN-α respectively, as a primary screen for identification of persons carrying neutralizing IFN-specific plasma Abs. First, MS patient plasma with a history of treatment with IFN-β were analyzed for anti-IFN-β activity. We identified two plasmas with ability to inhibit IFN-β-mediated activity (MS-001, and MS-006) in contrast to other MS plasma or non-MS control plasma M486-B55 or F623-C32 (Fig. 1A). Data from Fig. 1A only identified MS-001 and MS-006 as having suppressive activity as reflected when compared to other samples tested. Due to sample availability, we further tested MS-001 plasma inhibitory effect in the IFN-β dimerization assay using different concentrations of IFN-β (Fig. 1B top panel) or different concentrations of plasma and IFN-β (Fig. 1B bottom panel). MS-001 plasma showed no inhibition in the IFN-α dimerization assay even when lower concentrations of IFN-α were used together with the highest plasma concentration (11% final in the well) (Fig. 1C). To confirm that the observed activity was associated with anti-IFN-β IgG in plasma, sequential IgG depletions from MS-001 plasma were completed resulting in the loss of inhibitory effect in the IFN-β dimerization assay (Fig. 1D). Specifically, Fig. 1D and 1H top panels show western blots of 10% plasma at 5 consecutive times of anti-IFN-β and anti-IFN-α Abs depletion with Protein A/G sepharose beads. It should be noted that as shown in Supplemental Fig. 2, the depletion of Abs was specific as no depletion of albumin in the same plasma samples was observed. Note that remaining bands after deletion of H chain (55Kd) and L chain (25Kd) is likely non-specific binding of the rabbit anti-human IgG (H+L) primary Ab when used in human sera.

FIGURE 1.

Identification of neutralizing IFN-α-specific and IFN-β-specific plasma from IFN-α or IFN-β-treated persons. (A) Percent (%) inhibition of IFNAR dimerization in the presence of IFN-β and plasma from participants with MS (MS-001 to MS-011) or from healthy control individuals M486-B55, F623-C32. Plasma was pre-incubated overnight with IFN-β at 4°C, at 3X of the final concentrations (final concentrations in the wells with HEK IFNAR1/IFNAR2 cells: 11% plasma, 1.2 nM IFN-β). (B) Top panel shows luminescence for IFNAR dimerization by IFN-β (in limiting dilutions) in the presence or absence of 11% plasma from participant MS-001 or from healthy control individuals M486-B55 and F623-C32. Bottom panel shows luminescence for IFNAR dimerization by IFN-β (in limiting dilutions) in the presence or absence of plasma (in limiting dilutions) from participant MS-001. (C) Luminescence for IFNAR dimerization by IFN-α (in limiting dilutions) in the presence or absence of 11% plasma from participant MS-001 or from healthy control individual M486-B55. (D) Top panel shows Western blot of 10% plasma from participant MS-001 before (indicated as “neat”), and at five rounds of Ab depletion with A/G sepharose beads. Bottom panel shows % inhibition of IFNAR dimerization in the absence of 1.2nM IFN-β, or in the presence of 1.2nM IFN-β and 10% plasma from participant MS-001 (before and after Ab depletion) or from healthy control individual F623-C32. (E) Percent (%) inhibition of IFNAR dimerization in the presence of IFN-α and plasma from ART-suppressed persons living with HIV having a history of receiving IFN-α immunotherapy (Peg-002 to Peg-032), or plasma from healthy control individuals M486-B55, F623-C32, or neutralizing anti-IFN-α Ab (BMS-13H5)]. Plasma was heat inactivated at 56°C for 1 h and then pre-incubated overnight with IFN-α at 4°C, at 3X of the final concentrations (final concentrations in the wells with HEK IFNAR1/IFNAR2 cells: 11% plasma, 2.4 nM IFN-α, 1.1. μg/ml anti-IFN-α Ab). (F) Luminescence for IFNAR dimerization by IFN-α (in limiting dilutions) in the presence or absence of 11% plasma from participant Peg-008 or from healthy control individual F623-C32. (G) Luminescence for IFNAR dimerization by IFN-β (in limiting dilutions) in the presence or absence of 11% plasma from participant Peg-008 or from healthy control individual F623-C32. (H) Top panel shows Western blot of 10% plasma from participant MS-001 before (indicated as “neat”), and at 5 rounds of depletion of Abs with A/G sepharose beads. Bottom panel shows % inhibition of IFNAR dimerization in the absence of 2.4nM IFN-α, or in the presence of 2.4nM IFN-α and 10% plasma from participant Peg-008 (before and after Ab depletion) and from healthy control individual M486-B55. In panels (B), (C), (F) and (G) plasma was heat inactivated at 56°C for 1 h and then pre-incubated overnight with IFN-β or IFN-α at 4°C, at 3X of the final concentrations. Dilution values shown are the final concentrations in the wells with cells. Data in panels (A), (D), (E) and (H) are shown as means and SD of % inhibition. Experiments shown are representative of a minimum of duplicate experiments.

For activity against IFN-α among ART-suppressed persons living with HIV, only plasma from Peg-008 showed the highest inhibitory effect in the IFN-α dimerization assay even when lower concentrations of IFN-α were tested (Fig. 1E, F). No inhibitory effect was found against IFN-β in same assay by Peg-008 plasma (Fig. 1G). As above, activity against IFN-α in Peg-008 plasma was lost after Ab depletion from plasma (Fig. 1H). Overall, plasmas with IgG-mediated neutralizing anti-IFN-specific activity were identified.

Identification and cloning of sequence of diverse IFN-α and IFN-β binding Abs using pooled donor cDNAs into Ab phage display technology

To clone the neutralizing anti-IFN-α and IFN-β Abs identified above, a workflow based on Ab phage display technology from PBMC-cell cDNA as starting material was employed (Supplemental Fig. 3A). As described in the methods, total RNA was isolated from PBMC samples from MS-001 and Peg-008. RNA was pooled and a master cDNA was synthesized by reverse transcription. The variable region gene segments (VH and VL) of the heavy and light chains were then amplified via polymerase chain reaction. These VH and VL gene segments were subsequently cloned into specialized phagemid expression vectors, generating a library of single-chain Fv (scFv) DNAs, where “Fv” refers to the variable region fragment of an Ab. Through 2–4 rounds of panning against either IFN-α or IFN-β, remaining scFv/phage with the ability to selectively bind to a single IFN were identified. From bacterial colonies, individual scFv/phage (referred to as monoclonal phage) were isolated, enabling the identification of the top six specific binding clones (3 against IFN-α: AH07808, AH07856, AH07857, 3 against IFN-β: AH07859, AH07866, AH07882) and 5 clones with partial binding activity (AH07824, AH07831, AH07845, AH07858, AH07881).

The VH and VL DNA sequences of selected scFv/phages were subjected to sequencing for VH and VL fragment, and subsequently subcloned into expression plasmids (Fig. 2A) that incorporated Ab constant region domains, thereby reconstituting authentic bivalent IgG Abs. The nucleotide sequences of the top six anti-IFN-α2a and anti-IFN-β mAbs were analyzed for diversity by phylogenic trees based on the sequences of CDRs. Included in analysis were added partially reactive clones as well as nucleotide sequence of previously reported neutralizing Abs against IFN-α (BMS-13H5; patent W02005059106A2) or IFN-β (CTI-AF1; patent US20170313769A1) respectively (66, 67). Notably, diversification of CDRs was observed between the newly identified clones against IFN-α (Supplemental Fig. 3B) or anti-IFN-β (Supplemental Fig. 3C) and previously reported neutralizing Abs. This study presents a novel method for the identification of Abs selectively binding to IFN-α or IFN-β. Leveraging a multiple-persons pooled cDNA approach, clones with V regions of both kappa (VK) and lambda (VL) chains were isolated. These clones demonstrated efficacy in selectively binding to IFN-α or IFN-β mAbs. The results underscore the effectiveness of this approach in cloning diverse and specific CDRs targeting IFN-α and IFN-β. This method offers advantages in capturing a wide range of Ab specificities and can expedite the discovery and development of therapeutic Abs for various clinical applications.

FIGURE 2.

Characterization of expressed and purified mAbs using Western blot and Coomassie blue straining together with analysis of mAb reactivities with IFN-α2a and IFN-β proteins using ELISA and Western blot assays. (A) The amplified VH and VL genes were selectively assembled into DNA cassettes. The pairs of IgH and IgL genes were co-transfected in Expi293F cell culture. (B) Following expression, mAb were purified, and the purified IgG was subjected to SDS-PAGE analysis using a NuPAGE 4–12% Bis-Tris gel. Gel analysis revealed the presence of two bands in each lane, with approximate molecular weights of 25 kDa and 50 kDa, respectively. These bands corresponded to the light and heavy chains of the mAb and subjected to immunoblotting using the fusion proteins as indicated. To further confirm the identity of the heavy and light chains, Western blotting results displayed the heavy and light chains on the left side of the panel (lane P is positive control Human IgG 1K, lane 1 sample prepared under reducing conditions, lane 2 sample prepared under non-reducing conditions), confirming the successful expression and purification of the mAb. (C) For ELISA, purified Abs were diluted by serial dilution and incubated with purified IFN-α2a and IFN-β proteins. The reactivity of the purified Abs was determined by indirect ELISA, where the Ab titer was defined as the highest dilution of serum exhibiting an OD450 ratio. The results showed that the purified Abs exhibited high reactivity. The data are presented as the mean ± SD, with OD representing optical density. (D) In Western blot analysis, recombinant proteins (0.5 μg) were separated by SDS-PAGE, and equal concentrations of purified mAbs were used to probe the blots for purification and identification. A protein molecular weight marker (Lane M) was included. The primary Ab (at a dilution of 1:200) was recombinantly expressed with high specificity and affinity for the target proteins. The secondary Ab used was IRDye© 800 CW goat anti-human IgG (at a dilution of 1:10000). Western blotting confirmed that the observed band corresponded to the target proteins. Clones specific for IFN-α2a and IFN-α are indicated in orange fonts, while clones specific for IFN-β are indicated in purple fonts. Experiments shown are representative of three experiments.

Recombinant IgG production: Generation and characterization of IgG Abs

Sequenced DNA sequences described above encoding IgG heavy and light chains were synthetically constructed and inserted into the protein expression vector, pCDNA3.4, featuring a robust promoter (Fig. 2A). Notably, the immediate early CMV promoter was employed to enhance mRNA stability and translation efficiency within the expression vector (69). Ab production was conducted in Expi293F cell culture, resulting in the successful expression of eight IgGs (6 target clones, BS-13H5, and CTI-AF1). Furthermore, LALA mutations were introduced to the AH07808, AH07824, AH07856, AH07857, AH07859, and AH07866 clones for additional validation. Purification of the recombinant proteins was achieved using a Ni-NTA agarose resin column, following the manufacturer’s instructions. Following expression, mAb were purified, and the purified IgG was subjected to SDS-PAGE analysis using a NuPAGE 4–12% Bis-Tris gel. The recombinant proteins expressed displayed a molecular weight of approximately 150 kDa under non-reducing conditions (Fig. 2B, lane 2). Under denaturing conditions, the gel analysis revealed the presence of two bands in each lane with approximate molecular weights of 25 kDa and 50 kDa, corresponding to the light and heavy chains of the mAb. Confirmation of the heavy and light chains was also obtained by Western blotting to establish the successful expression and purification of each mAb (Fig. 2B, lane 1).

The reactivity of expressed mAbs against recombinant IFN-α2a and IFN-β proteins was then evaluated using ELISA (Fig. 2C) and Western blot analysis (Fig. 2D). Binding results demonstrated that clones AH07856 and AH07857 specifically reacted with IFN-α2a over AH07808, and all exhibited no cross-reactivity with IFN-β. Conversely, clones AH07859, AH07866, and AH07882 showed reactivity towards IFN-β over IFN-a by ELISA assays with minor detection of reactivity with IFN-a by Western for only AH07866 and AH07882. Added clones were generated (AH07824, AH07831, AH07845, AH07858, AH07881) yet showed binding to both IFN-a and IFN-b by either assay. Collectively, our findings confirm that human clones AH07856, AH07857, and AH07808 specifically recognize IFN-α2a without binding to IFN-β, while clones AH07859, AH07866, and AH07882 selectively recognize IFN-β with minimal to no cross-reactivity towards IFN-α2a.

Identification of clones with neutralization anti-IFN-specific activity against either IFN-α- or IFN-β-mediated dimerization of the IFN receptor

We determined the neutralization potential of the top 10 identified clones (i.e. AH07808, AH07824, AH07831, AH07845, AH07856, AH07857, AH07858, AH07859, AH07866, AH07882) when evaluated for inhibition of IFN-α or IFN-β-mediated dimerization of the IFN receptor.

Positive control conditions using described neutralization Abs against IFN-α (BMS-13H5) and IFN-β (CTI-AF1) reconfirmed sensitivity of assay to detect IFN-specific inhibition (Fig. 3A). Importantly, clones AH07856 and AH07857 inhibited IFN-α but not IFN-β, while clones AH07859 and AH07866 inhibited IFN-β but not IFN-α. These results were also confirmed for clones when the LALA mutation was introduced (data not shown).

FIGURE 3.

IFN-specific neutralization and binding saturation to human, rhesus or mouse IFN-α and IFN-β for anti-IFN-specific clones. (A) Percent (%) inhibition of IFNAR dimerization in the presence of IFN-α (top panel) or IFN-β (bottom panel) when combined with supernatants from IFN-specific binding clones, control plasma (MS-006 or Peg-008 from Fig. 1, respectively), or neutralizing control anti-IFN-α (BMS-13H5) and anti-IFN-β (CTI-AF1) Abs. Supernatants from clones or anti-IFN-α or anti-IFN-β Abs were pre-incubated overnight with IFN-α or IFN-β at 4°C at the final concentrations. Final concentrations in the wells with HEK IFNAR1/IFNAR2 cells: 67% clone’s supernatants, 2.4 nM IFN-α, 1.2 nM IFN-β, 10 μg/ml of anti-IFN-α (BMS-13H5) and anti-IFN-β (CTI-AF1) Abs. (B) Saturation binding of different concentrations of neutralizing anti-IFN-specific clones AH07856, AH07857, AH07859 and AH07866 to human IFN-α (left) or human IFN-β (right) panels. (C) Left column shows saturation binding of different concentrations of anti-IFN-α-specific clones AH07856 (top) and AH07857 (bottom) to human, rhesus, and mouse IFN-α with K_D values indicated next to each clone. Right column shows saturation binding of different concentrations of anti-IFN-β-specific clones AH07859 (top) and AH07866 (bottom) to human, rhesus, and mouse IFN-β with K_D values indicated next to each clone. See Table I and Supplemental Fig. 4 for added kinetic parameters for IFN-α and IFN-β clones binding to human, mouse and rhesus IFN-α and IFN-β, respectively. In panel (A) IFN-α and IFN-α-specific clones are indicated with orange fonts and orange shaded boxes respectively, while IFN-β and IFN-β-specific clones are indicated with purple fonts and purples shaded boxes respectively. In panels (B) and (C) saturation binding of different concentrations of clones to IFN-α and IFN-β from the different species are indicated with different shapes, while IFN-α and IFN-α-specific clones are indicated with orange fonts, and IFN-β and IFN-β-specific clones are indicated with purple fonts. Experiments shown are representative of a minimum of duplicate experiments.

Assessment of the binding affinity of the cloned anti-IFN-α and anti-IFN-β to human, rhesus and mouse IFN-α or IFN-β

To further explore the dimerization assay results, clones AH07856, AH07857, AH07859, and AH07866 were tested for their ability to bind either human IFN-α or IFN-β by SPR. The binding results for clones AH07856 and AH07857 are summarized in Fig. 3B and Table I. Briefly, clones AH07856 and AH07857 bound to IFN-α with nanomolar potency. However, no binding was detected by AH07859 or AH07866. By contrast, clones AH07859 and AH07866 were able to bind IFN-β, but no binding was detected for AH07856 or AH07857.

Table I.

Binding kinetics of clones AH07856, AH07857, AH07859 and AH07866 with human, rhesus and mouse IFN-α2a and IFN-β analyzed by SPR

	AH07856					AH07857						AH07859					AH07866
IFN-α	kon × 103 M−1s−1	koff × 10−4 s−1	t1/2 min−1	KD nM	Rmax RU	kon × 103 M−1s−1	koff × 10−4 s−1	t1/2 min−1	KD nM	Rmax RU	IFN-β	kon × 103 M−1s−1	koff × 10−4 s−1	t1/2 min−1	KD nM	Rmax RU	kon × 103 M−1s−1	koff × 10−4 s−1	t1/2 min−1	KD nM	Rmax RU
Human	410 ± 3.4	2.1 ± 0.18	55	0.52	263	1800 ± 13	6.9 ± 0.088	17	0.39	110	Human	4.2 ± 0.42	0.95 ± 0.088	122	23	233	2.3 ± 0.23	0.51 ± 0.19	227	22	53
Rhesus	510 ± 4.3	4.2 ± 0.24	28	0.82	205	250 ± 2.0	5.9 ± 0.18	20	2.4	111	Rhesus	5.3 ± 0.39	0.76 ± 0.076	152	14	220	0.4 ± 0.83	0.25 ± 0..18	462	62	276
Mouse	850 ± 8.6	15 ± 0.24	7.7	1.8	14	480 ± 6.6	20 ± 0.31	5.8	4.3	9.1	Mouse	No binding	No binding	No binding	No binding	No binding	No binding	No binding	No binding	No binding	No binding

Detailed kinetic studies were then carried out on these clones for their ability to bind human, rhesus, or mouse IFN-α or IFN-β. Data are summarized in Supplemental Fig. 4 showing binding kinetics of these clones over time for different concentrations of human, rhesus, or mouse IFN-α or IFN-β, and Fig. 3C showing the saturation binding and K_D values of these different concentrations of human, rhesus, or mouse IFN-α or IFN-β, while detailed binding kinetics of this saturation binding are shown in Table I. Briefly, the IFN-α clone AH07856 was able to bind both human and rhesus IFN-α with similar affinity (K_D values of 0.52 nM and 0.82 nM respectively), and also mouse IFN-α with slightly reduced affinity at 1.8 nM. The dissociation rate for AH07856 from mouse IFN-α was considerably faster compared to human and rhesus (Fig. 3C, Supplemental Fig. 4, Table I). Alternatively, the IFN-α clone AH07857 had much higher affinity for the human IFN-α (K_D = 0.39 nM) compared to rhesus IFN-α (K_D = 2.4 nM), or mouse IFN-α (K_D = 4.3 nM). The dissociation rate for AH07857 from the human IFN-α was considerably slower compared to the rhesus and mouse IFN-α (Fig. 3C, Supplemental Fig. 4, Table I). By contrast, the IFN-β clones AH07859 and AH07866 were only able to bind human and rhesus IFN-β, with no observable binding to mouse IFN-β. The affinities for these mAbs to IFN-β were considerably less than what was observed with the IFN-α mAbs, with K_D values of 23 nM and 14 nM to human and rhesus IFN-β, respectively for AH07859, and 22 nM and 62 nM to human and rhesus IFN-β for AH07866 (Fig. 3C, Supplemental Fig. 4, Table I). A weaker initial affinity for both clones against IFN-β is likely the result of a slow association rate together with a slower, once bound, dissociation rate than what was observed for the IFN-α mAbs.

Inhibition of IFN-mediated STAT1 phosphorylation in human cells by neutralizing anti-IFN-specific clones

To determine the ability of anti-IFN-specific clones AH07856, AH07857, AH07859 and AH07866 Abs to neutralize IFN-α or IFN-β signaling in primary lymphocytes, we first tested the induction of STAT-1 phosphorylation in primary CD4⁺ T cells by 1250 U/ml IFN-α or IFN-β in the presence of limiting dilutions of each cloned Ab starting at 25 μg/ml. Second, we tested the activity of each Ab at 25 μg/ml against decreasing amounts of IFN-α or IFN-β stimulation.

As expected by prior data, clones AH07856 and AH07857 strongly inhibited IFN-α but not IFN-β-mediated STAT-1 phosphorylation in CD4⁺ T cells (Fig. 4A, Supplemental Fig. 5). Inhibition of pSTAT-1 phosphorylation was further confirmed for both anti-IFN-α-specific clones at all decreasing IFN-α concentrations tested. (Supplemental Fig. 6).

FIGURE 4.

Inhibition of IFN-mediated STAT-1 phosphorylation in CD3⁺CD4⁺ primary T cells by anti-IFN-specific clones. Levels of phosphorylated STAT-1 (pSTAT-1) are shown in CD3⁺CD4⁺ T cells following stimulation of PBMC with or without IFN-α or IFN-β and in the presence or absence of anti-IFN-specific clones. (A) pSTAT-1 histogram panels for the effects of clones AH07856, or AH07857 (tested at 25, 5 and 1 μg/ml) on 1250 U/ml IFN-α or IFN-β stimulation. (B) pSTAT-1 histogram panels for the effects of clones AH7859 and AH07866 (tested at 25 μg/ml) on decreasing concentrations of IFN-α or IFN-β stimulation (1250, 625, 312, 156, 78 U/ml). For all histograms: (i) pink shaded peak shows constitutive pSTAT-1 levels in the absence of stimulation with IFN-α or IFN-β and in the absence of clones, (ii) green peak shows pSTAT-1 levels in the presence of stimulation with IFN-α or IFN-β and in the absence of clones, (iii) blue peak shows pSTAT-1 in CD3⁺CD4⁺ T cells in the presence of stimulation with IFN-α or IFN-β and in the presence of clones, and (iv) black line and number with red font inside the histogram show CD3⁺CD4⁺pSTAT-1⁺ percent (%) of CD3⁺CD4⁺ T cells following stimulation with IFN-α or IFN-β and in the presence of clones. IFN-α and IFN-α-specific clones are indicated with orange fonts, while IFN-β and IFN-β-specific clones are indicated with purple fonts. Experiments shown are representative of a minimum of duplicate experiments. See Supplemental Fig. 1 for gating strategy for the detection of pSTAT-1 in CD3⁺CD4⁺ T cells by flow cytometry.

On the other hand, clones AH07859 and AH07866 when tested at 25 μg/ml showed greater inhibition at lower concentrations of IFN-β (156 and 78U/ml) in contrast to partial inhibition observed at highest concentration of IFN-β tested (1250 U/ml) indicating greater potency with lower IFN-β levels (Fig. 4B). Interestingly, usage in the assay of decreasing concentration of clones AH07859 and AH07866 following stimulation with 312 U/ml or lower IFN-α showed partial inhibition of IFN-α-mediated STAT-1 phosphorylation whereas IFN-β-mediated STAT-1 phosphorylation was more efficiently inhibited (Fig. 4B). Overall, we confirm anti-IFN-specific inhibition of pSTAT-1 phosphorylation in primary cells exposed to either IFN-α-specific and anti-IFN-β-specific clones.

Specificity of neutralizing anti-IFN-specific clones in inhibiting IFN-mediated gene expression in human myeloid cells following exposure to IFN-α subtypes or IFN-β

As STAT-1 phosphorylation is an acute inhibition effect, we further tested the ability of these Ab clones to inhibit pSTAT1 phosphorylation (30 mins after stimulation) and expression of IFN-induced genes MX1 and USP18 (24 hs later) in PMA-differentiated THP-1 cells. We first titrated the inhibitory activity of the four anti-IFN-specific Abs when measuring MX1 induction by Western analysis at 24 hs after IFN stimulation (Fig. 5A). As before, we document IFN-specific inhibition of MX1 induction by all clones with greater potency by anti-IFN-α clones following IFN-α2a stimulation when compared to anti-IFN-β clones stimulated with IFN-β1a. Second, we tested by titration and breath the potency of the anti-IFN-α-specific clones when tested against 13 available IFN-α subtypes (Supplemental Table II) in addition to IFN-β1a. Both anti-IFN-α-specific clones effectively inhibited STAT-1 phosphorylation, MX1 and USP-18 by IFN-α2a, IFN-α2b, and IFN-αK (IFN-α6) with no inhibition activity against other IFN-α subtypes tested or IFN-β1a (Fig. 5B, Supplemental Fig. 7). Regarding anti-IFN-β-specific clones, we tested titrations of each clone (AH07859 and AH07866) or their combination against IFN-β1a-induced pSTAT-1, MX1 and UPS18 expression as above (Fig. 5C–D). Similar inhibition was observed by either Ab used singly or by their combination confirming decreased IFN-β-stimulated ISG expression. Taken together, data supports that each identified neutralizing IFN-specific Ab against exogenous IFN-α or IFN-β can inhibit activation of ISG induction in T cells and in myeloid cells.

Inhibition of type I IFN paracrine effects following LPS or S100A14 stimulation by neutralizing anti-IFN-specific Abs

To assess the neutralizing activity of the cloned anti-IFN-α-specific or anti-IFN-β-specific Abs against cells producing type IFNs following TLR stimulation, we tested the paracrine effects of TLR-4-induced type I IFN expression by LPS or S100A14 stimulation of PMA-differentiated THP1 cells. As shown in the top panel of Fig. 6, STAT-1 phosphorylation was achieved after 30 mins stimulation of the cells with exogenously added IFN-α or IFN-β (Fig. 6 sections C and D) with inhibition by clone AH07856 (Fig. 6 section C lane 1) and AH07866 (Fig. 6 section D lane 2) respectively. As expected, no detectable STAT-1 phosphorylation was observed at 30 mins after stimulation with LPS (Fig. 6 section A) or S100A14 (Fig. 6 section B) due to absence of signaling at this time otherwise present by induction of a prevalent IFN-β signature by 24 hs (67) as evidenced by the induction ISG proteins MX1 and USP18. Evidence for the inhibition of paracrine effects of IFN-b in mediating ISG expression are indicated by the decreased levels of MX1 and USP18 observed at 24 hs in the presence of neutralizing anti-IFN-β-specific (Fig. 6 sections A and B lane 2) over anti-IFN-α-specific Ab (Fig. 6 sections A and B lane 1). Importantly, the strength of neutralizing anti-IFN-α-specific and the neutralization activity of anti-IFN-β-specific Abs support their biological regulation potential.

FIGURE 6.

Inhibition of LPS or S100A14-induced IFN-mediated expression of MX1 and USP18 in THP-1 cells by neutralizing anti-IFN-specific Ab. Western blots showing the expression of pSTAT-1, STAT-1, MX1, USP18 and β-actin in THP-1 cells following stimulation for 30 mins for pSTAT-1 and STAT-1, and for 24 hs for MX1 and USP18 with IFN-α2a (0.01 nM, section C), IFN-β1a (0.01 nM, section D), LPS (10 ng/ml, section A) or S100A14 (10 ng/ml, section B) or unstimulated control (section E) in the presence of 30 μg/ml of clones AH07856 (lane 1) or AH07866 (lane 2) or in the absence of clones (lane 3). pSTAT-1/STAT-1: 91, 84 kDa, MX1: 76 kDa, USP18: 39, 34 kDa, and β-actin: 42 kDa. IFN-α and IFN-α-specific clones are indicated with orange fonts, while IFN-β and IFN-β-specific clones are indicated with purple fonts. Experiments shown are representative of three experiments.

Discussion

In this study we report novel natural human cloned neutralizing anti-IFN-specific Abs against IFN-α or IFN-β, able to modulate IFN-mediated signaling and gene expression. Importantly, we report success employing a phage-display strategy for cloning Abs using a pooled approach of cDNA from isolated total PBMC from multiple donors receiving IFN-α or IFN-β immunotherapy against HIV or to treat MS respectively, with pre-identified plasma Abs able to selectively neutralize IFN-α or IFN-β. By contrast to what was done in this study, conventional approaches to clone human mAbs are fraught with technical difficulties as they require the immortalization of human B cells or relay in sorting isolated antigen-specific B cells, and thus are not suitable for developing large panels of potentially therapeutic mAbs from target patients at once. Our data support that this strategy can be beneficial under the following conditions: (a) identification of patients with confirmed target plasma IgG-mediated activity, (b) availability of recombinant antigens for selective panning for phage selection, and (c) established functional assays able to evaluate and select target clones. The clinical impact of developing human neutralizing IFN-specific Abs against IFN-α and IFN-β is already evidenced by the clinical testing of the available humanized murine Abs targeting these molecules, as well as by the larger set of studies blocking all type I IFN signaling by directly targeting the IFNR (66, 70, 71). We now add two newly characterized natural human neutralizing anti-IFN-α-specific (AH07856, AH07857) and two anti-IFN-β-specific AH07859 and AH07866) Abs allowing for both basic and clinical studies targeting one or both human IFNs. Interestingly, we document that both neutralizing anti-IFN-α-specific Abs can target IFN-α2a and IFN-α2b as the major subtypes produced by myeloid and dendritic cells, as well as the IFN-αK subtype produced by keratinocytes and implicated in cutaneous lupus erythematosus and skin photosensitivity (72, 73).

A clear greater potency of the natural neutralizing anti-IFN-α-specific Abs over the anti-IFN-β-specific Abs cloned was detected in various in vitro response assays using recombinant IFNs (rIFN) proteins. We interpret these differences to be related to greater in vitro efficiency of IFN-β to bind the IFNR over IFN-α in presence of Abs, or a weaker slower on-rate interaction by cloned anti-IFN-β Abs when compared to those against IFN-α as suggested by kinetics of saturation binding data. Importantly, these differences did not prevent these clones from acting to neutralize endogenously produced IFN-β levels following TLR activation as evidenced by lower ISG induction 24hs later. Indeed, we establish the activity of anti-IFN-β-specific clone AH07866 to result in a greater inhibition of natural IFN-mediated effects in THP-1 cells following stimulation by either LPS, a major component of Gram-negative bacterial cell wall and a potent immunostimulatory product, or S100A14, a danger-associated molecular pattern and mediator of immune response (74). Both LPS and S100A14 are known to signal through Toll like receptor 4 (TLR4) (75) resulting in the production of type I IFNs (76). It is of interest to reconcile the “lower potency” interpretation of the anti-IFN-β Ab when tested with rIFN-β protein levels versus TLR- or S100A14 cell-stimulated levels of natural IFN-β suggesting rIFN levels used in biochemical or in vitro response assays may still represent an over-estimate of the physiologic target levels secreted by a cell. Our data also showed limited cross-reactivity by anti-IFN-β-specific Abs against IFN-α at lower concentrations as evaluated at STAT-1 induction 30 mins after stimulation, but without similar inhibition of ELISA binding, IFNR dimerization inhibition, MX1 induction 24 hours later, or detectable strength of binding affinity by SPR analysis, supporting the interpretation of a predominance of activity against IFN-β specificity over IFN-α.

Although our in vitro experiments document activity by all four Abs described, future studies will be needed to test their activity in vivo in order to reconfirm their potential to influence systemic IFN-mediated responses as predicted by our in vitro data. Regarding future in vivo testing, SPR analysis showed that all four neutralizing anti-IFN-specific clones had the ability to bind rhesus IFNs raising the possibility of testing a rhesus Ab version in this this model.

Type I IFNs have been shown to play a significant role in acute and chronic immune response and their regulation is evident in several clinical settings. As a result, the development of tools such as the Abs described here would allow to advance our understanding of type I IFN responses in cancer, autoimmune and infectious diseases (1, 77, 78) as exemplified by HIV infection where type I IFNs although antiviral have also been linked with lowering T cell mediated control of viral replication (49–51, 56).

In conclusion, we present a novel approach for cloning neutralizing Abs against IFN-α and IFN-β from pooled PBMCs. Utilizing molecular and immunological techniques, variable gene transcripts for both heavy (VH) and light (VL) chains of IgGs were amplified. Two pairs of potent natural mAbs capable of selectively neutralizing either IFN-α or IFN-β were identified. These Abs hold promise for therapeutic strategies targeting autoimmune diseases, viral infections, and other medical applications. Further research is essential to fully elucidate the therapeutic potential of these Abs in preclinical and clinical settings.

Key points.

• Cloning of neutralizing anti-IFN-α and IFN-β Abs from PBMCs

• Identification of two pairs of natural mAbs capable of neutralizing IFN-α or IFN-β

• Potential usage of the identified mAbs for therapeutic strategies

Abbreviations used in this article:

ISG

IFN-stimulated genes

MX1

MX Dynamin Like GTPase 1

USP18

Ubiquitin specific peptidase 18

MS

multiple sclerosis

ART

antiretroviral therapy

HEK

Human embryonic kidney

RT

room temperature

OBB

Odyssey Blocking Buffer

S100A14

S100 calcium binding protein A14

BCA

Bicinchoninic acid