Critical role of bioanalytical strategies in investigation of clinical PK observations, a Phase I case study

Abstract

RG7652 is a human immunoglobulin 1 (IgG1) monoclonal antibody (mAb) targeting proprotein convertase subtilisin/kexin type 9 (PCSK9) and is designed for the treatment of hypercholesterolemia. A target-binding enzyme-linked immunosorbent assay (ELISA) was developed to measure RG7652 levels in human serum in a Phase I study. Although target-binding assay formats are generally used to quantify free therapeutic, the actual therapeutic species being measured are affected by assay conditions, such as sample dilution and incubation time, and levels of soluble target in the samples. Therefore, in the presence of high concentrations of circulating target, the choice of reagents and assay conditions can have a significant effect on the observed pharmacokinetic (PK) profiles. Phase I RG7652 PK analysis using the ELISA data resulted in a nonlinear dose normalized exposure. An investigation was conducted to characterize the ELISA to determine whether the assay format and reagents may have contributed to the PK observation. In addition, to confirm the ELISA results, a second orthogonal method, liquid chromatography tandem mass spectrometry (LC-MS/MS) using a signature peptide as surrogate, was developed and implemented. A subset of PK samples, randomly selected from half of the subjects in the 6 single ascending dose (SAD) cohorts in the Phase I clinical study, was analyzed with the LC-MS/MS assay, and the data were found to be comparable to the ELISA data. This paper illustrates the importance of reagent characterization, as well as the benefits of using an orthogonal approach to eliminate bioanalytical contributions when encountering unexpected observations.

Keywords: 5, 5′-tetramethylbenzidine;; BSA, bovine serum albumin; CDR, complementarity-determining region; ELISA, enzyme-linked immunosorbent assay; HRP, horseradish peroxidase; IS, internal standard; IgG1, immunoglobulin G1; LC-MS/MS; LC-MS/MS, liquid chromatography tandem mass spectrometry; LDL-c, low density lipoprotein cholesterol; LDLR, low density lipoprotein receptor; LLOQ, lower limit of quantification; MAD, multiple-ascending dose; MQC, minimum quantifiable concentration; MRM, multiple reaction monitoring; NHS, normal human sera; PBS, phosphate buffered saline; PCSK9, proprotein convertase subtilisin/kexin type 9;; PD, pharmacodynamics; PK, pharmacokinetics; RG7652; RT, room temperature; S/N, signal-to-noise; SA, streptavidin; SAD, single-ascending dose; SIL, stable isotope-labeled; TMB, 3, 3′; clinical pharmacokinetic assay; enzyme-linked immunosorbent assay; mAbs, monoclonal antibodies; proprotein convertase subtilisin/kexin type 9; rhuPCSK9, recombinant human PCSK9; signature peptide.

Figure 1.

A target-binding ELISA PK assay was utilized to support RG7652 clinical Phase I study.

Figure 2.

An orthogonal method (LC-MS/MS assay) was utilized to confirm the RG7652 generated using the ELISA.

Figure 3.

Characterization of mAb 1.2A4 raised against the CDR of RG7652 (A) mAb 1.2A4 binds specifically to RG7652 (B) PCSK9 does not block mAb 1.2A4 binding to RG7652 (each data point represents the mean value of 3 runs).

Figure 4.

Effect of sample incubation time on the ELISA RG7652 PK assay. Percent recovery of RG7652 in 9 NHS samples spiked with 180 ng/mL (A) or 3000 ng/mL (B) of RG7652 at 2 hour versus overnight incubation (each data point represents the mean value of 3 runs). (C) Percent recovery of RG7652 in samples of 1 μg/mL RG7652 plus various levels of biotin-rhuPCSK9 at 2 hour vs. overnight incubation.

Figure 5.

Characterization of affinity capture LC-MS/MS assays. The assay using mAb 1.2A4 for capture has better sensitivity (A) and accuracy (B) in the presence of rhuPCSK9.

Figure 6.

RG7652 measurements of Phase I post dosing Day 29 samples from 150 mg (A) and 600 mg (B) cohorts using 2 affinity capture LC-MS/MS methods.

Figure 7.

Comparison of Phase I PK results generated using the ELISA and affinity capture LC-MS/MS (mAb 1.2A4 for capture) assays. (A) Similar results were observed in a subset of Phase I samples (n = 123) from all 6 SAD cohorts. (B) Overall good correlation was observed between the 2 sets of data (R² = 0.983).