The NISTmAb Reference Material 8671 value assignment, homogeneity, and stability

The NISTmAb Reference Material (RM) 8671 is intended to be an industry standard monoclonal antibody for pre-competitive harmonization of best practices and designing next generation characterization technologies for identity, quality, and stability testing. It must therefore embody the quality and characteristics of a typical biopharmaceutical product and be available long-term in a stable format with consistent product quality attributes. A stratified sampling and analysis plan using a series of qualified analytical and biophysical methods is described that assures RM 8671 meets these criteria. Results for the first three lots of RM 8671 highlight the consistency of material attributes with respect to size, charge, and identity. RM 8671 was verified to be homogeneous both within and between vialing lots, demonstrating the robustness of the lifecycle management plan. It was analyzed in concert with the in-house primary sample 8670 (PS 8670) to provide a historical link to this seminal material. RM 8671 was verified to be fit for its intended purpose as a technology innovation tool, external system suitability control, and cross-industry harmonization platform. Graphical abstract The NISTmAb Reference Material (RM) 8671 is intended to be an industry standard monoclonal antibody for pre-competitive harmonization of best practices and designing next generation characterization technologies for identity, quality, and stability testing. Electronic supplementary material The online version of this article (10.1007/s00216-017-0800-1) contains supplementary material, which is available to authorized users.


SUPPLEMENTARY MATERIALS AND METHODS
Formulation Buffer. Formulation buffer (12.5 mmol/L L-His, 12.5 mmol/L L-His HCl, pH 6.0) was utilized as the blank solution for measurements. Formulation buffer was prepared as follows: 1) 1.3129 g histidine monohydrochloride monohydrate and 0.9704 g L-histidine were weighed out and diluted in a beaker with ~450 mL type 1 deionized ultrafiltered water (DIUF) that was obtained from an ultrapure water system fed with reverse osmosis water and filtered through a 0.2 µm filter.; 2) while recording pH with a calibrated pH meter, pH was adjusted by drop-wise addition of 1 mol/L hydrochloric acid to 6.00±0.02; 3) the solution was transferred to a 500 mL volumetric flask, the beaker rinsed with DIUF water and the flask volume adjusted to 500.0 mL using the rinse water; 4) the solution was sterile filtered into a sterile plastic bottle using a 0.22 µm cellulose acetate membrane; 5) the buffer was stored at 2 °C to 8 °C.

NISTmAb PS 8670 and RM 8671 Samples. Vialing of NISTmAb PS 8670 and RM 8671
(lots 14HB-D-001, 14HB-D-002, and 14HB-D-003) is described in the first publication of this series [1]. Samples reserved from each RM 8671 lot for homogeneity and value assignment were selected as indicated in Table S1 and Fig. S1 below. Thaw/freeze and accelerated stability samples were pulled sequentially from rack 1 of lot 14HB-D-001. Samples remained frozen during the reservation step.  Table S2. to the uncertainty to photometric accuracy as a result of wavelength uncertainty (u w ) is also included. Traceability is ensured by including uncorrected bias in the measurement uncertainty rather than by bias correction [3,4]. The bias itself must therefore be added to the uncertainty in the decadic attenuance [3,4]. SRM 2031 utilized a coverage factor of 2; therefore, the combined standard uncertainty associated with the decadic attenuance is given by Equation S1 . (S1) The u D was calculated according to Equation S1, followed by propagation of error through Equation 1 to achieve the combined standard uncertainty associated with the concentration measurement (u c ) as given in Table S3. The Type B contributions to the combined standard uncertainty budget are included in Table S4. The cuvette path length was measured previously by the NIST Dimensional Metrology Group using a Coordinate Measuring Machine fitted with a fiber probe as well as using interference fringes in the near-infrared, resulting in an additional Type B uncertainty (u b ). The uncertainty contribution associated with the theoretical extinction coefficient has not been fully evaluated. Instead the extinction coefficient has been utilized here as a constant, accepted industry norm value. For that reason, the mass concentration is reported as a reference value rather than a certified value.  Value is reported in centimeters.

Physicochemical Reference and Informational Value Assignment and Homogeneity
Samples. A total of nine samples were reserved for physicochemical homogeneity from each lot as depicted in Table S1 and Fig. S1. Samples were selected from racks across the fill sequence.
The Combi dispenser (Thermo Scientific) described in [1] utilizes an 8 channel dispenser filling a given column simultaneously (e.g. column 1 rows A through H). Three samples from the same row in a given rack were selected in this matrix so that one sample from each row/rack# combination could be dispensed to individual physicochemical assays. They were stored at 80 °C.
One vial from each lot/rack (1, 50, and 90) was thawed at room temperature for a total of three vials per lot. Each 800 µL vial was aliquoted into 5 x 150 µL fractions into a clean Thermo test samples (no more than 10) -IQ -PS 8670 -Blank, and was repeated as necessary to analyze all samples. Instrument qualification and system suitability controls were required to pass method performance criteria as outlined during method qualification [5,6]. Replicates for each vial using CZE, CE-SDS, and SEC are as described in Table S5. The intended protocol for each lot was to perform triplicate repeatability on one vial (to obtain a measure of intra-vial repeatability) and individual measurements on two additional vials (to obtain a measure of intervial reproducibility) for physicochemical value assignment of each lot, as performed for CZE.
CE-SDS followed this protocol for 14HB-D-001; however, fewer injections were performed on 14HB-D-002 and 14HB-D-003 to allow analysis within 24 hours such that the sample remained stable in the autosampler under reducing conditions. SEC analysis is an orthogonal measure of size heterogeneity; therefore replicates were performed to model CE-SDS with one sequence initiated immediately upon thaw and a second sequence initiated after storage at 4 °C for 24 hours post thaw in order to span a similar analysis timeframe. To establish homogeneity and assign informational values, six vials (two obtained from The within day contributions ( ) to the variance cannot be independently measured as both factors are contributing to the within day measurement. Therefore, a one-way nested ANOVA is used to calculate contributions to the within day precision and the total within day inter-vial variation. Substituting ( ) we get: Where u 2 iv can be treated as a Type A uncertainty for RM 8671, as it was measured.
is a Type B uncertainty, as it is a reference value from previous qualification [5,6]. SEC, CE-SDS, and CZE uncertainty analyses were performed using this method to provide a combined standard uncertainty for RM 8671 (u c ) that incorporated PS method qualification experience and represents the intermediate precision at a level of one standard deviation.
Reference values and associated combined standard uncertainty for each physicochemical assay are given in Table S6 through Table S8.  Therefore additional Tables S9, S10, S11, and S12 are included in the ESM for PS 8670 (from the qualification exercise) and RM 8671 lots 14HB-D-001, 14HB-D-002, and 14HB-D-003 respectively.   Tables   Table S13 Subvisible particle concentrations (ECD ≥ 2   Thaw/Freeze Stability Samples. Intra-lot homogeneity analysis was performed as described in the main text (see Physicochemical Methods); therefore vials reserved for T/F stability were obtained from Rack 1 of 14HB-D-001 after homogeneity samples were removed.

Informational Value
All vials were thawed at room temperature, and thaw/freeze cycles were performed, with the samples evenly split between 20 °C and 80 °C as described in Table S15. These temperatures were selected for T/F stability to represent the range of temperatures commonly employed for long term storage of frozen solutions. After each Thaw cycle one vial was inverted five times to mix, divided into 150 µL aliquots, labeled as "XFZ_150 µL" and then frozen at 80 °C for use in CZE, CE-SDS, SEC, and peptide mapping. An additional five vials from each T/F cycle were labeled as "XFZ," frozen at 80 °C and retained for flow imaging and DLS. All samples were then refrozen. After each refreeze, the labeled vials remained in the frozen state and unlabeled vials were thawed in the next round. Therefore the label indicates the total number of thaw/freeze cycles the sample underwent beyond the necessary initial thaw. The same sequence was performed simultaneously on 15 PS 8670 vials at each temperature (with the exception of three vials removed at each round) for a direct comparison during flow imaging analysis. All vials were stored at 80 °C after completion of the thaw/freeze cycles. The reference value (UV-Vis, CZE, SEC, and nrCE-SDS) or informational value (FI and DLS) for lot 14HB-D-001 was used as the zero T/F control.
Accelerated Stability Samples. Intra-lot homogeneity analysis was performed as described in the main text (see Physicochemical Methods); therefore vials reserved for accelerated stability were obtained from Rack 1 and Rack 2 of 14HB-D-001 after homogeneity samples were removed. At the appropriate sample start time, the indicated number of vials (Table S16) was thawed at room temperature for thirty (30) min and inverted five times to mix.
They were then labeled "X Temperature_XXDays" to indicate the designated temperature condition and total exposure time and placed directly at the indicated temperature. The samples remained at the indicated accelerated stability temperature until Day 28.