Qualification of NISTmAb charge heterogeneity control assays

The NISTmAb is a monoclonal antibody Reference Material from the National Institute of Standards and Technology; it is a class-representative IgG1κ intended serve as a pre-competitive platform for harmonization and technology development in the biopharmaceutical industry. The publication series of which this paper is a part describes NIST’s overall control strategy to ensure NISTmAb quality and availability over its lifecycle. In this paper, the development and qualification of methods for monitoring NISTmAb charge heterogeneity are described. Capillary zone electrophoresis (CZE) and capillary isoelectric focusing (CIEF) assays were optimized and evaluated as candidate assays for NISTmAb quality control. CIEF was found to be suitable as a structural characterization assay yielding information on the apparent pI of the NISTmAb. CZE was found to be better suited for routine monitoring of NISTmAb charge heterogeneity and was qualified for this purpose. This paper is intended to provide relevant details of NIST’s charge heterogeneity control strategy to facilitate implementation of the NISTmAb as a test molecule in the end user’s laboratory. Graphical Abstract Representative capillary zone electropherogram of the NIST monoclonal antibody (NISTmAb). The NISTmAb is a publicly available research tool intended to facilitate advancement of biopharmaceutical analytics. Electronic supplementary material The online version of this article (10.1007/s00216-017-0816-6) contains supplementary material, which is available to authorized users.

High pH Stress Buffer. Phosphate buffer (50 mmol/L, pH 8.9) for high pH stress experiments was prepared as follows: 1) Weighed out 0.8924 g sodium phosphate dibasic dihydrate; 2) dissolved in ≈45 mL DIUF water; 3) adjusted pH to 8.9 using 0.1 mol/L sodium hydroxide and 0.1 mol/L hydrochloric acid; 4) quantitatively transferred to a 50 mL class A volumetric flask; 5) adjusted volume to 50 mL using DIUF water; 6) capped flask and inverted 10x to mix contents; 7) sterilized by passing through a 0.22 µm filter; 8) checked pH and adjust to 8.89 ± 0.02. Stored at room temperature.
Preparation of Stressed Samples. Portions of PS 8670 were buffer exchanged into one of the following buffers using Zeba 10K MWCO spin columns: formulation buffer; 50 mmol/L acetate, pH 3.7; or 50 mmol/L phosphate, pH 8.9. Samples were transferred to tightly capped 0.5 mL LoBind tubes and incubated at 40 °C for 8 d. A control sample in histidine buffer was stored at -80 °C until analysis. On day 8, samples were removed from the incubator and placed at -20 °C until analysis.

Capillary Isoelectric Focusing (CIEF)
CIEF Linearity/LOD/LOQ Sample Preparation. Samples for linearity/LOD/LOQ studies were prepared in a dilution series as follows. Master mix was prepared as in Table 1 according to the optimized method but scaled 23 fold. The master mix was vortexed 3 x 30 s on high before use. A master mix blank (3 reps) was prepared by mixing 54 µL of formulation buffer (12.5 mmol/L L-histidine, 12.5 mmol/L L-histidine HCl, pH 6.0) with 846 µL of master mix. A 0.6 mg/mL PS 8670 in master mix stock (in triplicate) was prepared by mixing 54 µL of 10 mg/mL PS 8670 in formulation buffer with 846 µL of master mix. A dilution series was prepared in triplicate from these solutions according to Table S1.  (Table S13). Retention times and corrected peak areas were recorded as in Table S2. Calculated parameters were derived from the recorded data using the following equations. Main group relative abundance was calculated according to equation S1: where CA main group is the corrected area of the main group, CA 2K is the corrected area of the 2K basic variant peak, CA 1K is the corrected area of the 1K basic variant peak, and CA acidic group is the corrected area of the acidic group as calculated by the 32Karat software. The relative abundance (RA) of acidic variants was calculated according to equation S2: The relative abundance of basic variants was calculated using equation S3: were calculated for each fit. The rSD was calculated using Equation S5.
where Y calc is the theoretical Y value calculated from the line of best fit, Y meas is the measured Y value, and n is the number of data points in the curve. The rrSD was calculated using Equation

Limit of Detection and Limit of Quantitation Calculations. Limit of Detection (LOD) and
Limit of Quantitation (LOQ) were calculated as follows. The 6-sigmal signal-to-noise (SNR) ratio and relative abundance (RA) of the 1K basic variant were recorded in triplicate at the target loading concentration (0.4 mg/mL). The RA 1K was calculated as in equation S7.
where CA is the time-corrected area determined by peak integration in 32 Karat. This peak was chosen because it exhibited SNR values in the range of 3-15 at the target loading concentration.
The LOD was calculated according to equation S8: where SNR 1K is the 6-sigma signal-to-noise ratio for the minor variant used for LOD calculation, taken from the 32Karat software, C inj is the concentration of total protein loaded in the injection (0.4 mg/mL) and V inj is the injection volume (5.89 x 10 -4 mL). V inj was calculated in CE Expert where d is the capillary internal diameter in meters, Δp is the pressure drop across the capillary in Pascals, η is the buffer viscosity in Pascal-seconds, l is the total capillary length in meters, and t is the injection time in seconds. The LOQ was calculated as in equation S10. (S10) The mass-based LOD and LOQ were converted to % relative abundance as in equations S11 and S12.
(S11) (S12) In the case of the CE assays discussed herein, C inj = C target because a minor variant (1K peak for cIEF and 2K peak for CZE) was present at appropriate SNR for this type of determination.  The effect of these threshold values on integration of PS 8670 peaks is detailed in Figure   S2. At a threshold of 100, both baseline and mAb peaks are integrated. Manual sifting of peaks is required to differentiate peaks of interest, which would introduce uncertainty between analysts with different judgements of what constitutes a mAb peak. At a threshold of 1000, a small amount of area attributable to the mAb is missed in the integration, but no manual data analysis is required. Table S3 gives the relative abundances of the charge variants at the different threshold levels determined from triplicate sample preparations analyzed on one day by one analyst. Differences in measured charge heterogeneity values are due largely to differences in integration of the acidic group, reflecting poor separation efficiency and high background in this region. Major technology development beyond the scope of this work would be necessary to achieve high quality separation of the many charge variants comprising the acidic group.

Discussion of Thresholding in CIEF Data
Barring this, we have erred on the side of conservative, automated integration in the interest of reproducibility and robustness over the lifetime of the product and method. The chosen integration parameters with a threshold level of 1000 are given in Table S13.
The buffer was prepared as follows: 1) dissolved 26.234 g EACA and 149 µL TETA in ≈450 mL DIUF) water; 2) adjusted the pH to 5.7 using glacial acetic acid; 3) transferred the   (Table S19 and S20). The integration parameters for the linearity/LOD/LOQ studies (Table S19) differ from the integration parameters used during qualification (Table S20) only in that the "Integration Off" command is set from 0 min to 8.5 min in the former and from 0 min to 4.0 min in the latter, to allow integration of the formulation buffer peak so that its retention time could be easily monitored.
Retention times and corrected peak areas were recorded as in Table S5. Main group RA, basic variants RA, and acidic variants RA were calculated using equations S1, S2, and S3 above.

Statistics for CZE including intermediate precision were calculated in Microsoft Excel
using the Analyse-it® plug-in (Analyse-it Software, Ltd., Leeds, UK) as discussed in [1].
Briefly, the precision for a given quality parameter was calculated by performing an ANOVA to estimate the total variance of the dataset and to model the components of the variance due to within-day variability (repeatability) and between-day variability (encompassing multiple columns, instrument drift, etc). This analysis was accomplished using the Analyse-it® measurement system analysis (MSA) Precision tool and setting the model to "Y with 1 random factor", where the factor was the date of analysis. The estimator was set to be standard deviation with a two-sided 95 % confidence interval. The method was chosen to be "Exact/MLS", and the "ANOVA" option was checked. Figure 10 in the main text) were prepared, except containing each individual data point (as opposed to averages) to allow a statistical fit evaluation. The plots were fit to linear regressions and statistical evaluations performed as described above for cIEF. The linear regression results for CZE data are described in Table S6. Triplicate injection of each PS 8670 dilution (9 injections per day) was performed on three different capillaries on six different days as outlined in Table S7. A fresh preparation of BGE was used with each capillary. The injection sequence for each day followed the form:

Summary of Qualified CZE Method
CZE Method Performance Criteria. The IQ standard and PS 8670 will be used to evaluate system performance and system suitability, respectively, during NISTmAb RM 8671 value assignment [2]. Performance criteria for the method were set for each parameter based on the measured intermediate precision. These criteria are useful for ensuring that the analytical method is in control, thus establishing confidence in the data acquired using the method. The criteria for the IQ are as follows:  Visually conforms to expectation (expected peak shape and pattern). The criteria for injections of PS 8670 are as follows:  Visually conforms to expectation (expected peak shape, no new peaks above LOD).

Capillary Day Number of PS 8670 Injections
Capillary # 1 1 9  The migration time of the main peak falls within ±3u c of the mean: (9.16 min to 10.19 min).
Blank injections should be performed at the start and end of each sequence and should contain no new peaks above the LOD. Fig. S1