Toxicological and pharmacological assessment of AGEN1884, a novel human IgG1 anti-CTLA-4 antibody

Abstract

CTLA-4 and CD28 exemplify a co-inhibitory and co-stimulatory signaling axis that dynamically sculpts the interaction of antigen-specific T cells with antigen-presenting cells. Anti-CTLA-4 antibodies enhance tumor-specific immunity through a variety of mechanisms including: blockade of CD80 or CD86 binding to CTLA-4, repressing regulatory T cell function and selective elimination of intratumoral regulatory T cells via an Fcγ receptor-dependent mechanism. AGEN1884 is a novel IgG1 antibody targeting CTLA-4. It potently enhanced antigen-specific T cell responsiveness that could be potentiated in combination with other immunomodulatory antibodies. AGEN1884 was well-tolerated in non-human primates and enhanced vaccine-mediated antigen-specific immunity. AGEN1884 combined effectively with PD-1 blockade to elicit a T cell proliferative response in the periphery. Interestingly, an IgG2 variant of AGEN1884 revealed distinct functional differences that may have implications for optimal dosing regimens in patients. Taken together, the pharmacological properties of AGEN1884 support its clinical investigation as a single therapeutic and combination agent.

Conflict of interest statement

Competing Interests: I have read the journal's policy and the authors of this manuscript have the following competing interests: Randi B. Gombos, Ana Gonzalez, Mariana Manrique, Dhan Chand, David Savitsky, Benjamin Morin, Ekaterina Breous-Nystrom, Christopher Dupont, Rebecca A. Ward, Mark A. Findeis, Jeremy D. Waight, Dennis Underwood, Christopher Clarke, Marc van Dijk, Jennifer S. Buell, Jean-Marie Cuillerot, Robert Stein, Elise E. Drouin and Nicholas S. Wilson have ownership of equity securities and/or are currently employed by Agenus Inc. or a subsidiary thereof. This does not alter our adherence to PLOS ONE policies on sharing data and materials.

Figures

Fig 1. AGEN1884 binds CTLA-4 and blocks CTLA-4 from interacting with CD80 and CD86.

(A) Binding of fluorescently-labeled CD80-Fc or CD86-Fc (1 nM) in the presence of increasing concentrations of AGEN1884 or an IgG1 isotype control. Binding to CTLA-4-linked microspheres was assessed using Luminex. (B) CTLA-4-expressing CHO cells were pre-incubated with increasing concentrations of AGEN1884 or an IgG1 isotype control followed by the addition of a fixed concentration of fluorescently-labeled CD80-Fc or CD86-Fc (0.625 μg/ml). Binding of CD80-Fc or CD86-Fc to the CHO cells was assessed by flow cytometry. (C-D) AGEN1884 binding to a (C) Jurkat cell line genetically engineered to express human CTLA-4 or (D) wildtype (CTLA-4-negative) Jurkat cell line. Expression of CD28 was also assessed using an anti-CD28 antibody (empty histogram) compared to an isotype control (filled histogram) on the (E) CTLA-4-expressing and (F) wildtype (CTLA-4-negative) cells lines. (A-D) Representative data from one of three experiments indicate the mean ± SEM in each treatment group.

Fig 2. AGEN1884 epitope.

(A) Mapping by hydrogen-deuterium exchange mass spectrometry (HDX-MS) on the CTLA-4 structure (PDB 1I8L). Ribbon representation of CTLA-4 highlighting residues identified as having reduced HDX by HDX-MS: residues 80–82 (QVT, magenta), 135–139 (YPPPY, cyan), 140–141 (YL, dark blue), 142–149 (GIGNGTQI, pale blue). (B) Structure of the human co-stimulatory complex CD80 (grey)/CTLA-4 (yellow) (PDB 1I8L). Residues having reduced HDX are indicated as in panel (A). The view of the structure highlights the loop region comprised of residues at positions 135–149, which encompasses a turn-loop that directly interacts with CD80. (C) Sequence alignment between human and cynomolgus macaque CTLA-4. An asterisk indicates identical residues, a colon indicates conservation between groups of strongly similar properties and a period indicates conservation between groups of weakly similar properties. Solid line boxes indicate residues identified as having reduced HDX.

Fig 3. AGEN1884 increases T cell activation alone and in combination with other immunomodulatory antibodies.

(A) A Jurkat T cell line genetically engineered to express CTLA-4 with an IL-2-dependent luciferase reporter was co-cultured with CD80/CD86-positive artificial APCs in the presence of increasing concentrations of AGEN1884 or an isotype control antibody. The fold increase in relative light units (RLU) relative to baseline is shown. (B) CTLA-4 expression was measured on T cells on days 1, 2, 4, 8 and 10 following stimulation with a sub-maximal concentration of SEA peptide (100 ng/mL). Primary human PBMC were stimulated with a sub-maximal concentration of the SEA peptide (100 ng/mL) and (C) increasing doses of AGEN1884, (D) single dose of AGEN1884 (10 μg/mL), (E) AGEN1884 (10 mg/mL) ± nivolumab (anti-PD-1 antagonist antibody), pembrolizumab (anti-PD-1 antagonist antibody), 25F7 (anti-LAG-3 antagonist antibody), 10C7 (anti-CD137 agonist antibody) (10 mg/mL) or (F) AGEN1884 (10 mg/mL) ± AGEN2034 (anti-PD-1 antagonist antibody) (10 mg/mL). Replicate cell supernatants were collected after 5 days for measurement of IL-2. Representative data from one of three experiments indicate the mean ± SEM in each treatment group (n = 3). (D-F) Data were analyzed using a Student’s t-test. Significant differences depicted were p

Fig 4. Fcγ receptor signaling and antibody-dependent cellular cytotoxicity of CTLA-4-expressing cells with AGEN1884.

(A-C) Jurkat cell lines genetically engineered to express FcγRs upstream of a NFAT-dependent luciferase reporter were co-cultured with CTLA-4-expressing cells and increasing doses of AGEN1884 or an isotype control. Signaling through (A) FcγRIIIA-V158-, (B) FcγRIIIA-F158-, or (C) FcγRIIA-H131-expressing cells was assessed based upon luciferase expression, which is shown as relative light units (RLU). (D) The cytotoxicity of CTLA-4high Jurkat cells by CD16-expressing NK-92 cells based on lactate dehydrogenase (LDH) release in the presence of increasing concentrations of AGEN1884 or an isotype control antibody. (E-F) Extracellular CTLA-4 expression (empty histogram) compared to the isotype control (filled histogram) was measured on (E) CD3+FoxP3+ or (F) CD3+FoxP3- cells. The lysis of (G) CTLA-4highCD3+FoxP3+ target cells or (H) CTLA-4lowCD3+FoxP3- target cells by primary NK cells in the presence of increasing concentrations of AGEN1884 compared to an isotype control. Representative data from one of three experiments indicate the mean ± SEM in each treatment group (n = 2–3). (G, H) Data were analyzed using a Student’s t-test for each dose of AGEN1884 compared to the isotype. Significant differences depicted were p<0.05 (*).

Fig 5. Pharmacokinetic profile of AGEN1884 and T cell-dependent antibody response and immune cell response to systemic co-administration of AGEN1884 with an ENGERIX-B® and KLH vaccine in cynomolgus macaques.

AGEN1884 titers were measured from serum samples collected from five groups of animals (n = 3 per group) following (A) a single intravenous (IV) dose at either 3 mg/kg or 100 mg/kg or (B) five weekly IV doses at 5 mg/kg, 30 mg/kg, or 100 mg/kg. Cynomolgus macaques (n = 6 per group) were given 10 mg/kg of AGEN1884 via IV administration with an ENGERIX-B® and KLH vaccine on days 1 and 29. Duplicate samples were analyzed for (C) anti-HBsAg-specific IgG serum titers or (D) the frequency of IFN-γ-producing HBsAg-specific cells in PBMC based on the number of spot forming units (SFU) per million PBMC. Anti-KLH (E) IgG and (F) IgM serum titers were also measured.

Fig 6. AGEN1884 in combination with anti-PD-1 further potentiates T cell proliferation in vivo.

Cynomolgus macaques (n = 6 per group) were administered 3 mg/kg of nivolumab alone or in combination with 10 mg/kg of AGEN1884 via intravenous (IV) infusion. Duplicate samples of PBMC were analyzed for Ki67 expression in CD4+ and CD8+ T cells at 7 days prior to treatment or 3, 10, 14, 18 and 22 days after treatment. (A-B) Representative dot plots of CD4+ and CD8+ T cells isolated from PBMC from cynomolgus macaques given (A) 3 mg/kg of nivolumab alone or (B) a combination of 10 mg/kg of AGEN1884 with 3 mg/kg of nivolumab. (C) Representative dot plots of naïve, central memory and effector memory CD4+ and CD8+ T cells isolated from PBMC from cynomolgus macaques given a combination of 10 mg/kg of AGEN1884 with 3 mg/kg of nivolumab.

Fig 7. Fcγ receptor signaling and antibody-dependent cellular cytotoxicity of CTLA-4-expressing cells with AGEN2041 and increased T cell activation in the presence of AGEN2041.

(A-C) Jurkat cell lines genetically engineered to express FcγRs upstream of a NFAT-dependent luciferase reporter were co-cultured with CTLA-4-expressing cells and increasing doses of AGEN2041 or an isotype control. Signaling through (A) FcγRIIIA-V158-, (B) FcγRIIIA-F158-, or (C) FcγRIIA-H131-expressing cells was assessed based upon luciferase expression which is shown as relative light units (RLU). (D) The cytotoxicity of CTLA-4high Jurkat cells by CD16-expressing NK-92 cells based on lactate dehydrogenase (LDH) release in the presence of increasing concentrations of AGEN2041, AGEN1884 (positive control), or an IgG2 isotype control antibody. (E-F) Primary human PBMC were stimulated with a sub-maximal concentration of the SEA peptide (100 ng/mL) and increasing doses of (E) AGEN2041 versus (F) AGEN1884. Cell supernatants were collected after 5 days for measurement of IL-2. Representative data indicate the mean ± SEM in each treatment group (n = ≥2).