Pharmacokinetics and pharmacodynamic effect of crenezumab on plasma and cerebrospinal fluid beta-amyloid in patients with mild-to-moderate Alzheimer's disease

Abstract

Background: Crenezumab, a fully humanized anti-beta-amyloid (Aβ) immunoglobulin G4 (IgG4) monoclonal antibody, binds to both monomeric and aggregated forms of Aβ. We assessed the pharmacokinetics (PK)/pharmacodynamics (PD) of crenezumab and its interaction with monomeric Aβ(1-40) and Aβ(1-42) peptides in serum/plasma and cerebrospinal fluid (CSF) samples from the phase II ABBY and BLAZE studies and the phase Ib GN29632 study.

Methods: In ABBY, BLAZE, and GN29632 studies, patients with mild-to-moderate AD were treated with either placebo or crenezumab (300 mg subcutaneously every 2 weeks [q2w], or 15 mg/kg, 30 mg/kg, 45 mg/kg, 60 mg/kg, or 120 mg/kg intravenously q4w). Serum/plasma PK/PD analyses included samples from 131 patients who received crenezumab in all three studies. CSF PK/PD analyses included samples from 76 patients who received crenezumab in ABBY or BLAZE. The impact of baseline patient factors on Aβ profiles was also evaluated.

Results: The serum concentration of crenezumab increased in a dose-proportional manner between 15 and 120 mg/kg q4w. Total monomeric plasma Aβ(1-40) and Aβ(1-42) levels significantly increased after crenezumab administration. The mean crenezumab CSF to serum ratio was ~ 0.3% and was similar across dosing cohorts/routes of administration. No clear correlation was observed between crenezumab concentration and Aβ(1-42) increase in CSF at week 69. The target-mediated drug disposition (TMDD) model described the observed plasma concentration-time profiles of crenezumab and Aβ well. Elimination clearance (CLel) and central volume of distribution (Vcent) of crenezumab were estimated at 0.159 L/day and 2.89 L, respectively, corresponding to a half-life of ~ 20 days. Subcutaneous bioavailability was estimated at 66.2%.

Conclusions: Crenezumab PK was dose proportional up to 120 mg/kg, with a half-life consistent with IgG monoclonal antibodies. Our findings provide evidence for peripheral target engagement in patients with mild-to-moderate AD. The study also showed that a model-based approach is useful in making inference on PK/PD relationship with unmeasured species such as free plasma Aβ levels.

Trial registrations: ABBY: ClinicalTrials.gov, NCT01343966. Registered April 28, 2011.

Blaze: ClinicalTrials.gov, NCT01397578. Registered July 19, 2011. GN29632: ClinicalTrials.gov, NCT02353598. Registered February 3, 2015.

Keywords: Nonlinear mixed-effects modeling; PK/PD; Pharmacokinetics; Plasma beta-amyloid; Target-mediated drug disposition model.

Conflict of interest statement

KY, AM, TB, SO, HL, LH, JYJ, and AQ are full-time employees of F. Hoffmann-La Roche Ltd/Genentech, Inc., and hold stocks in F. Hoffmann-La Roche Ltd/Genentech, Inc.

Figures

Fig. 1

Schematic representation of the TMDD model structure. According to Michaelis–Menten approximation, the drug–target complex is expected to be in quasi-equilibrium with the concentration of monoclonal antibody > > target concentration. Abbreviations: AAβ amount of beta-amyloid, Aβ beta-amyloid, Ccren concentration of crenezumab, IV intravenous, Kd equilibrium constant governing antibody-ligand binding, kdeg first-order rate constant for free ligand degradation, kint zero-order input rate constant for ligand, SC subcutaneous, TMDD target-mediated drug disposition

Fig. 2

Mean (SD) serum crenezumab concentrations after initial dose (weeks 1–5). Each line represents mean crenezumab serum concentration following IV administration; doses of 15 mg/kg (ABBY; BLAZE) and 30–120 mg/kg (GN29632 phase Ib study). Abbreviations: IV intravenous, SD standard deviation

Fig. 3

Mean (SD) total plasma Aβ(1–40) and Aβ(1–42) concentrations after initial dose (weeks 1–5). Total number of patients included = 72. Each line represents mean total Aβ(1–40) or Aβ(1–42) plasma concentration following IV administration; data shown are from the phase II ABBY study (SRI cohort) for 15 mg/kg dose and phase Ib GN29632 for 30–120 mg/kg doses. Abbreviations: Aβ beta-amyloid, IV intravenous, SD standard deviation, SRI safety run-in

Fig. 4

Serum crenezumab and total plasma Aβ Ctrough concentrations. Total number of patients included = 130. Circles represent data from the phase II ABBY and BLAZE studies for all patients for 300 mg q2w SC and 15 mg/kg q4w IV, and from the phase Ib GN29632 study for 30–120 mg/kg q4w IV from weeks 5, 9, and 13. Solid lines and shaded areas represent the median and 5–95% prediction intervals, respectively, based on target-mediated drug disposition model predictions. Abbreviations: Aβ beta-amyloid, Ctrough pre-dose serum crenezumab, IV intravenous, q2w every 2 weeks, q4w every 4 weeks, SC subcutaneous

Fig. 5

Steady-state pre-dose crenezumab concentration and Aβ(1–42) change from baseline in CSF at week 69. Data shown are from the phase II ABBY and BLAZE studies for patients enrolled in 300 mg q2w SC or 15 mg/kg q4w IV cohorts. Abbreviations: Aβ beta-amyloid, CSF cerebrospinal fluid, IV intravenous, q2w every 2 weeks, q4w every 4 weeks, SC subcutaneous

Fig. 6

Prediction-corrected visual predictive check plots. Phase Ib, data from GN29632 study; phase II, data from ABBY and BLAZE studies. Circles represent observed data. Solid and dashed lines represent median and 90th percentiles of observed data, respectively. Gray and blue shaded areas represent simulated 90% prediction interval of median and 90th percentiles, respectively. Abbreviations: Aβ beta-amyloid, Ph phase

Fig. 7

Predicted impact of patient baseline characteristics on PK profile of crenezumab and plasma Aβ levels. White vertical line refers to the predicted parameters for a 72-year-old male, weighing 72.3 kg, with a GFR of 72.5 mL/min/1.73 m2 after a crenezumab 60 mg/kg IV dose. Red bar depicts the 5th–95th percentile crenezumab exposure (left column) or Aβ levels (right column) range across the entire population. Green bar represents the influence of a single covariate on the predicted parameters. The upper and lower values for each covariate represent 90% of the observed covariate range in the population. Abbreviations: Aβ beta-amyloid, AUC area under the curve, BWT body weight, Cmax peak concentration, GFR glomerular filtration rate, IV intravenous, PK pharmacokinetics

Fig. 8

Simulations illustrating effects of varying crenezumab doses (mg/kg q4w) on plasma Aβ kinetics based on the developed PK/PD model. Note that total and free crenezumab concentrations overlap with each other. Abbreviations: Aβ beta-amyloid, PD pharmacodynamics, PK pharmacokinetics, q4w every 4 weeks