Mining the human autoantibody repertoire: isolation of potent IL17A-neutralizing monoclonal antibodies from a patient with thymoma

Abstract

Anti-cytokine autoantibodies have been widely reported to be present in human plasma, both in healthy subjects and in patients with underlying autoimmune conditions, such as autoimmune polyendocrinopathy candidiasis ectodermal dystrophy (APECED) or thymic epithelial neoplasms. While often asymptomatic, they can cause or facilitate a wide range of diseases including opportunistic infections. The potential therapeutic value of specific neutralizing anti-cytokine autoantibodies has not been thoroughly investigated. Here we used mammalian cell display to isolate IL17A-specific antibodies from a thymoma patient with proven high-titer autoantibodies against the same. We identified 3 distinct clonotypes that efficiently neutralized IL17A in a cell-based in vitro assay. Their potencies were comparable to those of known neutralizing antibodies, including 2, AIN457 (secukinumab) and ixekizumab that are currently in clinical development for the treatment of various inflammatory disorders. These data clearly demonstrate that the human autoantibody repertoire can be mined for antibodies with high therapeutic potential for clinical development.

Keywords: AIN457; APECED, autoimmune polyendocrinopathy candidiasis ectodermal dystrophy; CDR, complementary-determining region; CMC, Chronic mucocutaneous candidiasis; FACS, fluorescence-activated cell sorting; HFF-1, Human Foreskin Fibroblasts; IL17; IL17A, Interleukin 17A; PBMCs, peripheral blood mononuclear cells; RT-PCR, Reverse transcription polymerase chain reaction; Sindbis virus; huFc-γ1, human Fc-gamma 1; human autoantibodies; ixekizumab; mAb, monoclonal antibody; mammalian cell display; monoclonal antibodies; scFv-Fc; scFvs, single chain variable fragments; secukinumab.

Figure 1.

Detection of anti-IL17A autoantibodies and evaluation of their biological activity. (A) Particle-based multiplex determination of anticytokine autoantibodies for patient plasma (○) versus control plasma (boxes indicate healthy control values min to max with line at mean, N = 10). (B) HFF-1 cells were incubated with control (N = 5) or patient plasma and left unstimulated or stimulated with IL17A overnight. Culture supernatants were collected and evaluated for IL6 production. Shown is the fold-increase as the ratio of IL6 levels measured in stimulated vs. unstimulated culture supernatants. Average values represent data from 3 independent experiments.

Figure 2.

Neutralization of IL17A-induced IL6 expression by scFv-huFc-γ1 expressed in HEK-293T culture supernatants. HFF-1 cells were treated with 50 ng/ml human IL17A in the presence or absence of serial dilutions of the indicated antibodies expressed as huFc-γ1fusion proteins in HEK-293T culture supernatants. After 24 hours, supernatants were analyzed for IL17A-induced IL6 expression by IL6 sandwich ELISA, and the 50% inhibitory concentration (IC50) was determined. Benchmark antibodies are represented as dark bars. Error bars show 95% confidence intervals.

Figure 3.

Amino acid sequence alignments of variable light and variable heavy chains of anti-IL17A antibodies. (A) Alignment of variable light chain amino acid sequences of patient-derived IgG1 clones 5M002, 5M007, 5M010, 5M012, 5M024 and of benchmark antibodies AIN457 and ixekizumab (LY2439821). Sequences are numbered starting at the first amino acid of mature Ig light chain protein. Complementarity determining regions (CDRs) are indicated by brackets. Point mutations distinguishing clones 5M007 and 5M024 and clones 5M010 and 5M012 are indicated by arrows. Conservative and identical sequences are shown in gray. (B) Alignment of variable heavy chain amino acid sequences of clones 5M002, 5M007, 5M010 5M012, 5M024 and of benchmark antibodies AIN457 and ixekizumab. Sequences are numbered starting at the first amino acid of mature Ig heavy chain protein. CDRs are indicated by brackets. Conservative and identical sequences are shown in gray. Sequences were aligned using AlignX.

Figure 4.

Binding properties of the identified IL17A-specific fully human monoclonal antibodies. (A) ELISA-based comparison of IL17A binding of patient-derived antibodies 5M002, 5M007, 5M012, 5M024, benchmark antibody AIN457, and an irrelevant IgG1. Binding of antibodies (at concentrations ranging from 3000 pM to 4.12 pM) to recombinant human IL17A (2 μg/ml) is illustrated. EC50 values for IgG1 binding to IL17A were determined by GraphPad Prism using nonlinear regression (curve fit) and sigmoidal dose response parameters. (B) Specificity of anti-IL17A antibodies. Patient-derived anti-IL17A antibodies 5M002, 5M007, 5M012, 5M024, and benchmark antibody AIN457 were analyzed using a particle-based multiplex assay against 21 cytokine targets at a final concentration of 200 ng/ml.

Figure 5.

Neutralization of IL17A by fully human monoclonal anti-IL17A autoantibodies. (A) IL6 response measured at OD₄₀₅ upon stimulation of HFF-1 cells with 25 ng/ml IL17A in presence of serial dilutions of identified anti-IL17A autoantibodies and benchmark antibodies AIN457 and ixekizumab (10 μg/ml to 5.6 × 10⁻⁵ μg/ml). (B) Diagram of IC50 values (nM) determined by detection of IL6 secreted from HFF-1 cells upon stimulation with IL17A (25 ng/ml) in presence of serial dilutions (10 μg/ml to 5.6 × 10⁻⁵ μg/ml) of 5M002, 5M007, 5M012, and 5M024 (light bars) in comparison to the 2 benchmark antibodies AIN457 and ixekizumab (dark bars). IC50 values were calculated in GraphPad Prism using nonlinear regression (log inhibitor versus response) for analyses. Error bars show 95% confidence intervals.