(M0930-06-34) Development and Qualification of an ECLIA Method for the Detection of Fremanezumab in Human CSF

Purpose: To explore target engagement of an anti-CGRP monoclonal antibody, fremanezumab, in the central nerve system (CNS), a novel LBA method was developed with appropriate surrogate matrix for the quantification of fremanezumab in human Cerebrospinal fluid (CSF).

Methods: Design of Experiment (DOE) on JMP platform was applied for the selection of suitable antibody pair for a Meso Scale Discovery (MSD) sandwich assay to measure fremanezumab concentration in human CSF. Different combinations of 5 capture and 4 detection antibodies were evaluated using assay buffer (1% BSA, 0.1% Tween 20 in PBS). Once the best antibody pairs were chosen, to find the best surrogate matrix to CSF, commercial artificial CSF (aCSF, Tocris Bioscience) and Superblock T20 (SB, Thermo Fisher) both supplemented with BSA or CSF at different concentrations were used to prepare calibration standard (STD) and compared to assay buffer. Using the conditions predicted from the DOE experiment, each STD curve was used to evaluate the recovery of Quality Control (QC) samples prepared in 100% CSF and diluted at different MRDs. Assay conditions were further optimized, and the best pair of capture and detection antibodies was selected for the final format of the method. The method uses an MSD Gold 96‑well Streptavidin Sector plate blocked with SB before addition of the capture reagent, which is a biotin‑conjugated capture antibody. Each plate contains fremanezumab STDs, QCs, and test samples in duplicate. The STDs are prepared fresh daily in the selected surrogate matrix. The QCs and samples are in 100% CSF and stored frozen until testing. After thawing, QCs and samples are diluted at the selected MRD before adding to the plate. The detection reagent is a Sulfo‑TAG‑conjugated detection antibody. The concentration of fremanezumab in QC and test samples is interpolated from the standard curve, adjusted by the dilution factor if applicable, and averaged for each sample. The mean of the back‑calculated concentrations of the duplicates is the reportable value. The following parameters were assessed to qualify the method: accuracy and precision (A&P), selectivity, dilutional linearity and stability.

Results: After initial screening of 5 capture antibodies (C1-C5) and 4 detection antibodies (D1-D4) candidates, three antibody pairs (C1/D1, C1/D4, C5/D4) were selected based on signal to noise (S/N) ratio and assay range. Second round of DOE was used to optimize the concentration of capture and detection antibodies. The best performing antibody pairs, C1/D4 and C4/D5, were further selected. SB supplemented with 33% CSF appeared to be the best surrogate matrix to prepare the STD that led to good recoveries of QC samples tested at MRD of 1:3 in SB with C1/D4 paired antibodies. The ECLIA method for the quantification of fremanezumab in human CSF was qualified with assay range from 10 ng/mL to 400 ng/mL in SB supplemented with 33% CSF, and with 4PLMarquardt and 1/Y² weighting factor fitting model. The overall precision of the STDs was ≤2%; the accuracy varied from 91 to 114%. The sensitivity of the method is confirmed at 30 ng/mL of fremanezumab in 100% human CSF. The inter-assay precision of the QCs was ≤8% and the accuracy was between 90 to 117%. Dilutional linearity was demonstrated up to a 36-fold dilution factor without hook effect observed at the highest tested concentration of 4320 ng/mL. The method is selective in human CSF samples. The samples are stable after 3 freeze/thaw cycles, at both room temperature and at 2-8°C for 24 hours, and at less than ≤-65°C.

Conclusion: An ECLIA method to quantify fremanezumab in CSF was developed successfully by overcoming challenges during the method development. The method was demonstrated to be reproducible, accurate and selective. The qualified method was applied in a clinical study to help explore target engagement of fremanezumab to CGRP in the CSF.