MSACL 2023 Abstract

INTRODUCTION:
The current trend in the pharmaceutical industry to quantify therapeutic antibodies in plasma or serum samples from clinical studies involves bioassays such as enzyme-linked immunosorbent assay (ELISA). Typically, these methods have been restricted to quantification of a single analyte and require the antibody structure and large sample volume. Several literature research findings suggest the possibility of structural changes an IgG2 antibody might undergo. If true, it can diminish or compromise the quantitation results obtained from ELISA. On the contrary, mass spectrometry (MS) allows the detection of proteotypic peptides and is independent of the antibody structure. In this study, we developed and applied liquid chromatography-parallel reaction monitoring (LC-PRM)/MS assays to quantify the concentration of a therapeutic humanized IgG2 antibody for 15 human plasma samples and compared LC-PRM/MS and ELISA results.

OBJECTIVES:
The key objectives of this project were to (1) develop quantitative LC-PRM/MS assays targeting tryptic peptides unique to the therapeutic IgG2 antibody, establishing the most efficient and reproducible digestion conditions, as well as the linear ranges, and (2) to apply these assays to the analysis of 15 human patient plasma samples covering up to 8 cycles of treatment from a phase II study, and compare the PRM/MS and ELISA concentrations.

METHODS:
Heavy and light antibody chains were in-silico digested using trypsin. Following stringent selection criteria, we ended up with two proteotypic peptides for the light and heavy chains, respectively, all absent in the human proteome.

Light/natural (NAT) and 13C/15N-stable isotope-labelled standard (SIS) peptides were provided by our supplier (SynPeptide, China) and assessed in-house for accurate peptide concentrations and purities by amino acid analysis (AAA) and capillary zone electrophoresis (CZE), respectively.

Peptide precursors were optimized by loading 100 fmol per peptide onto EvoTips, followed by analysis on an Evosep One HPLC with the 60 sample per day (SPD) method (21-minute gradient, column EV1109) and a PRM detection method measuring 2+ and 3+ precursor charge states on a Thermo Q Exactive Plus Orbitrap.

After establishing detectability of the antibody peptides at 5,000 ng/mL in human plasma by PRM/MS, a time course digestion study was performed evaluating three different denaturants (urea, sodium deoxycholate, and trifluoroethanol), seven digestion time points (0, 0.5, 1, 2, 4, 6 and 18h), and spiking the internal standard peptides before or after digestion. The samples used were plasma spiked with 20,000 ng/mL target antibody. The ideal condition was selected based on plateaued light/heavy ratios across three replicate preparations per condition, coefficients of variation (CVs) below 15%, maximal light/heavy peak area ratio, and maximal light peak intensity.

Next, the linear ranges of the PRM/MS assays were assessed with plasma sample spiked with 500,000 ng/mL of antibody, serially diluted with pooled human plasma to 200 ng/mL for a total of 10 concentration levels plus blank.

Finally, the concentrations of the target peptides in 15 experimental samples provided by a biotechnology company were quantified and compared with the antibody concentrations obtained from ELISA.

RESULTS:
All four peptides were most intense with the 2+ precursors. In the time course digestion study, urea as the denaturant with SIS spiked before digestion produced the most reproducible concentrations with good accuracy (within 80%-120% range), and CVs <15% among the replicates across all time points. Using a 2-hour digestion protocol, the linear ranges were 200 to 500,000 ng/mL for three peptides, and 200 to 50,000 ng/mL for the fourth peptide, suggesting sufficient sensitivity and linear range for expected antibody concentrations. Finally, 15 plasma samples from a phase II study at different dosing time points of up to 148 days were analyzed with the final PRM/MS approach, and results were compared to ELISA. While the ELISA results indicated no accumulation of the target antibody in the patient plasma, the PRM/MS approach confirmed the expected antibody accumulation as the dosing regimen progressed, thus providing evidence that the ELISA approach falsely underrepresented the actual antibody concentration in plasma, likely due to structural changes of the antibody in plasma.

CONCLUSION:
Using a highly reproducible and sensitive LC-PRM/MS approach, we were able to confirm the accumulation of a therapeutic antibody in patient plasma over the duration of 148 days in a phase II trial, whereas ELISA failed to demonstrate this behaviour.