A Cell-Based Potency Assay for Determining the Relative Potency of Botulinum Neurotoxin A Preparations Using Manual and Semi-Automated Procedures

Abstract

Cell-based potency assays (CBPAs) are required for the potency testing and commercial release of botulinum neurotoxin (BoNT)-based drug products. These CBPAs must account for the toxin's biological activities while meeting regulatory guidelines for precision and accuracy. Here, studies describe the characterization and qualification of the BoSapient CBPA and demonstrate that it is fit for use as a relative potency assay for BoNT/A-containing samples. The CBPA is operated in a 96-well plate format and relies upon the fluorescence emissions of a reporter that directly responds to BoNT/A activity. The BoSapient cell line expresses the BoNT/A-receptors SV2 and complex gangliosides, is responsive only to intact BoNT/A, and can robustly detect picomolar and sub-picomolar BoNT/A quantities, making the CBPA appropriate for quantifying BoNT/A-based drug products. The cell line was passaged 30 times without significant loss of reporter expression or BoNT/A sensitivity. Manual and semi-automated CBPA methods were developed and qualified according to regulatory guidelines and shown to have low bias (<4% from expected) and high precision (standard deviation < 8) across all test concentrations. Furthermore, the semi-automated method using the CBPA is demonstrated to improve intermediate precision by 39% compared to the manual method, while reducing operator dependency during method execution.

Keywords: botulinum neurotoxin; cell-based potency assay; laboratory automation; parallel line analysis; relative potency.

Figure 1

The BoSapient A CBPA. (A) Overview of the BoSapient CBPA. The CBPA uses an engineered human cell line that stably expresses a protein reporter containing CFP (green circle) and YFP (yellow circle) connected by full-length SNAP-25 and targeted to the cellular membrane. Refer to the main text for additional details. (B) BoSapient cells show reduced YFP emissions but not CFP emissions following incubation with BoNT/A. Cells were incubated in SAM with or without 300 pM BoNT/A for 48 h before imaging using an InCell Analyzer microscope. (C) Fluorescent emissions of the BoSapient CBPA in response to BoNT/A. The CBPA was executed as described in Section 4 with the indicated BoNT/A concentration and using a fluorescent plate reader to capture YFP and CFP responses. Background-subtracted YFP (left) emissions are divided by the background-subtracted CFP (center) emissions for a given assay plate well to generate an emission ratio (right). The emission ratio values were plotted as a function of BoNT/A concentration and fitted using a 4-parameter logistic (4PL) regression. Error bars represent the standard deviation of the mean. Limits of detection and quantitation were calculated from the mean emission ratio of the negative control wells minus 3 × standard deviation (SD) or 6 × SD, respectively.

Figure 2

The receptor binding domain of BoNT/A is required to elicit a response in the BoSapient CBPA. Cells were fixed and stained against SV2 isoforms SV2A, SV2B, and SV2C (A) or GD1a and GT1b (B) and imaged as described in Section 4. (C) A CBPA was executed as described in Section 4 except that seeded assay plates were pre-treated without (control) or with 1 µM recombinant BoNT/A heavy-chain (HcR/A) in assay media for 1 h at 37 °C before treatment with serial dilutions of BoNT/A holotoxin for 48 h at 37 °C. Emission ratios are plotted as a function of BoNT/A concentration (top). Bright-field and fluorescent images are shown for representative wells treated with 0 or 100 pM BoNT/A (bottom). (D) The CBPA was executed as described in Section 4 except assay plates were treated with serial dilutions of either BoNT/A holotoxin or BoNT/A light-chain (LC/A) for 48 h at 37 °C. Emission ratios are plotted as a function of BoNT/A or LC/A concentration.

Figure 3

Passage stability of the BoSapient cell line. Three vials, filled at the beginning (3/300), middle (151/300), or end (297/300) of a 300-vial master cell bank, were thawed from liquid nitrogen and passaged every 3–4 days (see Section 4). Every 5 ± 2 passages, 4-day cultures were subjected to a CBPA, as described in Section 4. (A) Cell viability during passaging at the indicated passage number for cultures derived from the three vials. A lower specification limit (LSL) of 90% was applied to the study. (B) Cell yields for 3- (left) and 4-day (right) cultures at the indicated passage number for cultures derived from the three vials. LSL and upper specific limit (USL) are based on yields obtained from the cell line during development. (C) YFP signal-to-background from execution of CBPAs as a function of passage number for all three vials. YFP signal-to-background was calculated by dividing the YFP emissions from assay plate wells containing cells but no BoNT/A (negative control) by the YFP emissions from wells containing no cells (blank). LSL and USL are calculated from mean from all CBPA runs ± 20%. (D) CBPA curve depth as a function of passage number for all three vials. Curve depth was calculated by subtracting the emission ratio (assay response) at the highest BoNT/A concentration from the emission ratio at the lowest BoNT/A concentration. LSL and USL are set as the beginning curve depth (passage 6) ± 50%. (E) CBPA relative potency results as a function of passage number and vial.

Figure 4

Semi-automated and manual operation of relative potency methods using the BoSapient CBPA. (A) Final assay plate layout for both the manual and semi-automated methods. Reference sample (RS, pink), test sample 1 (TS1, green), and test sample 2 (TS2, blue) serial dilutions (i.e., Dil_1, Dil_2, Dil_3, Dil_4) are applied in quadruplicate (e.g., RS-1, RS-2, RS-3, and RS-4) with each replicate independently diluted in parallel. Wells containing no cells (Blank, orange) or cells with no BoNT/A (dark green) are included as controls. Replicates are arrayed to minimize plate effects. (B) Workflow of the semi-automated and manual CBPA methods. Operator-performed actions for each method are shown in boxes. Both methods share cell maintenance, assay plate seeding, and initial sample preparation/suspension procedures but diverge during sample serial dilution, assay plate washing, and sample application procedures; these actions are fully automated in the semi-automated method. Thus, the operator only needs to insert the assay plate into the platform when prompted. Operator-performed steps that diverge between the manual and semi-manual methods are highlighted in orange.

Figure 5

Qualification of semi-automated and manual BoSapient CBPA methods. CBPAs were executed as described in the Experimental Procedures section. Test samples were formulated to 64, 80, 100, 125, or 150% of a 4 pM nominal concentration, corresponding to 80–120% of a typical 80–125% drug product release specification. (A) Example CBPA runs from the semi-automated method qualification. Assay plates were run independently with different pipetting heads fitting the automated pipetting platform. The theoretical potency of the two test samples were 64 and 125% and each was run against a 100% reference. Emission ratio was plotted as a function of BoNT/A concentration, assuming 100% nominal concentrations. (B) Example CBPA runs from the manual method qualification. Assay plates contain test samples of the same theoretical potency as described in (A), but the assays were executed independently and manually by two different operators. (C) Relative potency results from the semi-automated method qualification plotted as a function of expected/theoretical concentration. Each data point represents the mean and SD of six independent determinations. Data were fit by linear regression. (D) Relative potency results from the manual method qualification plotted as a function of expected/theoretical concentration. Data were treated as described in (C).

Figure 6

Linearity and bias of the semi-automated and manual BoSapient CBPA methods. (A) Quadratic plots of the studentized residuals of all results of the manual (left) and semi-automated (right) method qualifications were constructed as a function of theoretical relative potency and used to establish the linear range of the methods. Red shading represents the limits of the 95% confidence interval (CI). A studentized residual limit of ±1.96 was used to determine the upper (UL) and lower (LL) limits of the assay’s linear range. (B) Accuracy or bias, the difference between the mean measured RP value and the expected RP value at each tested theoretical concentration, was determined for the manual (left) and semi-automated (right) method qualification data and plotted versus theoretical potency. Based on typical 80–125% drug product specification limits, a bias limit of ±9 was calculated (20% of tolerance). (C) Accuracy results and acceptance status for all tested concentrations in the two methods.

Figure 7

Predicted OOS rates for the manual (A) and semi-automated (B) BoSapient CBPA methods. To determine the out-of-specification (OOS) rate as a percentage or in parts per million (PPM), the mean including analytical bias and the associated method standard deviation was determined from the method qualification data. Applying lower (80%, LSL) and upper (125%, USL) specification limits, the area under the normal curve outside of the specification limits was then calculated using the following equations: %OOS = (1 − NORMSDIST((Mean − USL)/standard deviation) + NORMSDIST((Mean − LSL)/standard deviation)) × 100 and OOS rate PPM = (1 − NORMSDIST((Mean − USL)/standard deviation) + NORMSDIST((Mean − LSL)/standard deviation)) × 1 × 10⁶.

Figure 8

Testing of commercial drug product with the semi-automated BoSapient CBPA method: CBPAs were executed as described in Section 4. Test samples were formulated to 80 or 125% of a 30 U/mL (BOTOX) or 40 U/mL (Xeomin) nominal concentration. Reference samples were formulated to 100% of the nominal concentration. Example CBPA runs for BOTOX (A) and Xeomin (B) are shown. Emission ratio was plotted as a function of drug product concentration assuming 100% nominal concentrations. (C) Relative potency results from three replicate CBPA runs of the semi-automated method for both BOTOX and Xeomin.