Target engagement and cellular fate of otelixizumab: a repeat dose escalation study of an anti-CD3ε mAb in new-onset type 1 diabetes mellitus patients

Abstract

Aims: This paper describes the pharmacological findings from a study where otelixizumab, an anti-CD3ɛ mAb, was dosed in new onset Type 1 diabetes mellitus (NOT1DM) patients. This is the first time that the full dose-response of an anti-CD3ɛ mAb has been investigated in the clinic. The data have been validated using a previously developed pharmacokinetic/pharmacodynamic (PK/PD) model of otelixizumab to simulate the interplay between drug administration, CD3ɛ target engagement and downmodulation.

Methods: Patients were randomized to control or active treatment with otelixizumab (1:4), administered via infusion over 6 days, in a dose-ascending study consisted of three cohorts (n = 10 per cohort) at doses of 9, 18 or 27 mg respectively. The study allowed quantification of otelixizumab PK, CD3ɛ target engagement and its pharmacodynamic effect (CD3ε/TCR modulation on circulating T lymphocytes).

Results: Otelixizumab concentrations increased and averaged to 364.09 (54.3), 1625.55 (72.5) and 2781.35 (28.0) ng ml-1 (Geom.mean, %CV) at the 9, 18 and 27 mg dose respectively. CD3ɛ target engagement was found to be rapid (within the first 30 min), leading to a receptor occupancy of ~60% within 6 h of dosing in all three doses. A dose-response relationship was observed with the two highest doses achieving a ~90% target engagement and consequential CD3ɛ/TCR downmodulation by Day 6.

Conclusions: Data from this study revealed maximum target engagement and CD3ɛ/TCR modulation is achieved at doses of 18, 27 mg of otelixizumab. These findings can be useful in guiding dose selection in clinical trials with anti-CD3ɛ mAbs.

Trial registration: ClinicalTrials.gov NCT02000817.

Keywords: Otelixizumab; PK/PD; dose-response; immunoinflammation; target engagement; type 1 diabetes.

Conflict of interest statement

G.V., A.N., R.B., K.B., J.B., D.I., K.P., S.Z., C.O.S. and A.v.M. are employees and shareholders of GlaxoSmithKline. B.K. reports personal consulting fees from GlaxoSmithKline during the conduct of this study; grants from H2020 EU and JDRF. D.L. and M.P. are employees of Eurofins Pharma Bioanalysis Services UK Limited; Eurofins received fees from GlaxoSmithKline for aspects of this study's conduct. Q.L. reports personal fees from SGS Exprimo NV, during the conduct of the study.

Funding for this trial was provided by GlaxoSmithKline (NCT02000817).

© 2018 The British Pharmacological Society.

Figures

Figure 1

CD3ε target engagement and downmodulation on CD8+ T cells. Otelixizumab was dosed over six days in three separate cohorts at a total dose of 9 mg, 18 mg and 27 mg. PBMC were stained with anti‐CD8, anti‐CD4, anti‐CD3 and anti‐human IgG F (ab’)2 antibodies and analysed by flow cytometry. Based on relative antibody molecules bound to the cells, the percentage of occupied CD3ε binding sites during first 24 h (Panel A) and during 14‐day period are depicted (Panel B). The percentage of occupied and downmodulated CD3ε during 14‐day period is depicted (Panel C). All panels represent means and associated 90% confidence intervals

Figure 2

Otelixizumab‐mediated CD3/TCR downmodulation on CD8+ T cells is independent of Fc portion. PBMC were cultured for 24 h in the presence of either otelixizumab whole antibody or F (ab’)2 fragments lacking the Fc portion. Cells were then stained with anti‐CD8 and TCR antibodies. Data represent MESF values of TCR sites for each otelixizumab concentration (mean of three donors ± SD)

Figure 3

Otelixizumab‐induced downmodulation of TCRs inhibits antigen‐dependent T cell proliferation. PBMC from two different healthy volunteers were stimulated for 4 days with Revaxis antigens (inactivated vaccine containing diphtheria, tetanus and polio) in the absence (control) or presence of different concentrations of otelixizumab. Mean expression of TCR on CD4+ T lymphocytes (as a percentage of control) is depicted as a function of mean percentage inhibition of proliferation. Closed circles and corresponding values indicate tested concentrations of otelixizumab. To estimate fit, R2 coefficient was calculated by Microsoft Excel

Figure 4

Serum otelixizumab concentrations for all three dose levels. Graph represents geometric mean data and associated 90% confidence intervals. Horizontal line indicates lower limit of quantification

Figure 5

Fluorescence intensity (± STD of four healthy donors) of FITC‐otelixizumab in the membrane of CD8+ T cells decreases over the first 24 h, and then remains stable for the subsequent 24 h. Representative images of a single CD8+ T cell from the same donor at each time point (not the same cell over time) show this loss of fluorescence is primarily due to internalization of FITC‐otelixizumab (green), which occurs within 1 h (punctate spots of fluorescence in the cytoplasm). Concomitantly, binding of the competitive anti‐CD3ε antibody (orange) increases over the same time period. This antibody cannot internalize as the cells are exposed to it only after fixation, thus demarcating the cell membrane very well and demonstrates availability of free CD3ɛ. A minimum of 500 CD8+ T cells were analysed per sample

Figure 6

The percentage of CD8+/otelixizumab cells with measurable levels of co‐localization as defined by bright detail similarity score (two donors ± STD) increases over time. Images are representative of a single CD8+ T cell from one donor at each time point (not the same cell over time). Lysosomes (LAMP‐1 – pink) are clearly observed in the cytoplasm at 0 h with FITC‐otelixizumab (green) restricted to the cell membrane. After incubation at 37°C, FITC‐otelixizumab begins to internalize and where this co‐localizes with LAMP‐1 a composite yellow fluorescence is generated. At 1 h, some internalized FITC‐drug is not localized with a lysosome in this particular cell. A minimum of 500 CD8+ T cells were analysed per sample

Figure 7

Visual predictive check of change from baseline in observed and simulated data using the otelixizumab PK/PD model. The different treatment arms are represented in different colours. Solid lines represent the global median of simulations and the ribbons around them show the corresponding 90% confidence intervals of these median. Dashed lines represent the global 5th and the 95th percentiles of simulations and the areas around them show the corresponding 90% confidence intervals. Dots depict observed data from OTX116505 study