Interlaboratory Method Validation for Determination of Polysorbate 80 for Analysis of mAbs

Abstract

Polysorbate 80 (PS80) is a widely used surfactant in therapeutic monoclonal antibodies (mAbs) because it prevents mAb aggregation and stabilizes the drug product. Various analytical methods have been published and are available for measuring PS80, each with its own limitations. In this study, an HPLC method employing an evaporative light scattering detector (ELSD) was developed for multiproduct analysis of PS80. Method validation was conducted across multiple laboratories using different types of instruments and ELSD detectors. The method was validated using various immunoglobulin G (IgG) mAb molecules and different sources of PS80, with the results showing that it can accurately measure PS80 concentrations ranging from 0.05 to 0.5 mg/mL. Robustness tests demonstrated that the method tolerates up to 160 mg/mL mAb interference, through either sample dilution or protein precipitation. This method can be recommended as a release/quality control method for PS80 concentration determination in therapeutic mAbs.

INTRODUCTION

Polysorbate 80 (PS80) is a nonionic surfactant widely used as a pharmaceutical excipient in therapeutic monoclonal antibody (mAb) formulations.¹⁾ PS80 reduces inter-protein molecular interactions and protects mAb molecules from aggregation and denaturation,²⁾ thereby enhancing product stability during manufacturing, transportation, and storage.^3–5) However, high concentrations of PS80 may raise safety concerns such as hemolysis, while insufficient amounts offer less protection for product stability. Recent findings indicate that degraded PS80 releases free fatty acids, which can lead to particle aggregation.^6–10) Therefore, precise control and accurate measurement of PS80 content are essential in mAb formulations.

There are no chromophore groups in PS80, which makes it challenging to detect using conventional analytical detectors. As a result, common UV detectors are ineffective for this purpose. Several test methods are currently available to measure PS80 content, often involving modification of PS80 with chromophore groups, such as fluorescence micelle analysis (FMA), cobalt thiocyanate colorimetry, and Coomassie Brilliant Blue colorimetry. These indirect methods break down the PS80 molecule and measure the products via reversed-phase liquid chromatography (RPLC) or gas chromatography (GC). Volatile molecules are detected after separation by chromatographic columns coupled with nonselective detectors such as an evaporative light scattering detector (ELSD) and a charged aerosol detector (CAD).^11–18) Although cobalt thiocyanate colorimetry is included in the Chinese Pharmacopoeia, it is a time-consuming and labor-intensive method. The CAD detector has a relatively high resolution, but the corresponding protein or formulation buffer also causes significant interference with the detection. In addition, GC-MS methods have also been used to determine the content of PS80; however, their pretreatment process is complex and difficult to perform.¹⁹⁾ For therapeutic mAbs with low concentrations of PS80 or high concentrations of protein, it is often necessary to add a step to remove proteins, and the method development is also more difficult. Within the common concentration range of PS80 from 0.05 to 0.5 mg/mL, ELSD is more applicable and widely used. Therefore, our objective was to develop a state-of-the-art platform for PS80 quantification in mAb product formulations.

We identified a candidate method based on ELSD and optimized its parameters to accommodate various mAb (immunoglobulin G [IgG]) molecules and concentrations, as well as different HPLC systems and ELSD detectors. The method was validated across 4 laboratories using different types of instruments and detectors, following the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use—Validation of Analytical Procedures Q2 (ICH Q2) guidelines. This method enhances testing efficiency while providing appropriate accuracy and precision, significantly reducing corporate compliance burdens. It can be recommended as a release/quality control method for PS80 concentration determination in therapeutic mAbs.

MATERIALS AND METHODS

Reagents

We used PS80 reference material (from J. T. Baker, Radnor, PA, U.S.A. and Nanjing Well, Nanjing, China), ultrapure water, and various buffers, including the drug product formulation buffer (DP buffer; 20 mM histidine, 240 mM sucrose, pH 5.5 ± 0.5) and other buffers as follows: A1, 20 mM histidine/histidine hydrochloride, 250 mM sucrose, pH 6.0; A2, 20 mM Tris, 0.1 mM diethylene triamine penta-acetic acid (DTPA), 10% mannitol, 0.1 M sodium chloride; B1, 20 mM sodium citrate, 50 mM sodium chloride, 3.0% mannitol, 20 μM DTPA, pH 6.0; B2, 25 mM l-arginine hydrochloride, 20 mM histidine, 12.5 mM sodium acetate, 5% (w/v) sucrose, pH 5.9; and C, 10 mM l-histidine, 252 mM d-sorbitol, 10 mM l-methionine, pH 6.25. The mAbs used included mAb-1 (IgG1, pI = 7.4) at a concentration of 72 mg/mL and mAb-2 (IgG1, pI = 7.0, 180 mg/mL). Formic acid, isopropanol, and acetonitrile (HPLC grade; Sigma-Aldrich, St. Louis, MO, U.S.A.) were also used.

Instruments

The HPLC system was equipped with a flow switch valve, an ELSD detector, a chromatography data station, a nitrogen source with purity >99%, a Waters Oasis MAX column (2.1 × 20 mm, 30 μm; Waters Corporation, Milford, MA, U.S.A.), and either a nitrogen blower or a vacuum concentrator. Instruments and detectors used in each laboratory are shown in Table 1.

Table 1. Instruments and Detectors Used in Each Laboratory

Lab No.	LC system	ELSD	Data station
Lab 1	Waters e2695 HPLC	Alltech 2000 ES	Waters Empower
Lab 2	Agilent 1260 HPLC	Agilent G4260B ELSD	Agilent OpenLAB CDS ChemStation
Lab 3	Waters e2695 HPLC	Waters 2424 ELSD	Waters Empower
Lab 4	Agilent 1260 HPLC	Agilent G4260A ELSD	Agilent OpenLAB CDS ChemStation

METHOD PARAMETERS

Gradient Conditions

Mobile phase A consisted of 2% formic acid in water, and mobile phase B consisted of 2% formic acid in isopropanol. The sample temperature was maintained below 10°C, while the column temperature was controlled at 30 ± 2°C. The standard injection volume was typically 20 μL. However, the diluted sample described in “Sample dilution” required 100 μL to obtain the target amount of PS80. The sampling rate was set to low or slow in the software (e.g., the syringe draw rate [μL/s on the Waters HPLC] was set to slow).

The gradient was run at a flow rate of 1.25 mL/min. Initial conditions were set at 90% mobile phase A. Mobile phase B increased to 20% between 1.0 and 3.4 min, followed by a hold at 100% B from 3.5 to 4.6 min. The gradient returned to the initial conditions at 4.7 min and was held constant until 6.6 min.¹¹⁾ We recommend switching the flow direction from the waste path to the detector path at 2.4 min and then returning the path back to waste at 6.6 min.

Detector Settings

The Agilent ELSD parameters (Agilent Technologies, Santa Clara, CA, U.S.A.) were set as follows: gas flow (GAS) at 1.00 SLM, nebulizer temperature (NEB) at 45°C, and evaporation temperature (EVAP) at 100°C. The time constant (SMTH) was 30, and the data output rate (HZ) was 10 readings/s. Detector gain (PMT) and LED intensity were adjustable. The signal for the highest PS80 concentration (0.5 mg/mL) was controlled at approximately 800 mV to prevent exceeding the detector’s upper limit. For the Alltech 2000ES ELSD, the alternative parameters were as follows: gas flow at 3.00 L/min, tube temperature at 100°C, impactor off, and gain of 1. For the Waters ELSD detector, the gas pressure was set at 25 psi, the nebulizer operated in heating mode with a power level of 50%, the drift tube temperature was maintained at 95.0°C, the gain value was 250, the data acquisition rate was 10 pps, and the time constant was 1.0000 s.

PS80 Reference Material, PS80 Test Control, and Sample Preparation

Standard curves were prepared using PS80 stock solution A (2 mg/mL PS80 in water), which was diluted to concentrations of 0.05, 0.075, 0.15, 0.25, and 0.5 mg/mL in water. This 0.05–0.5 mg/mL range encompassed the typical PS80 levels found in most mAb formulations. A PS80 test control at 0.15 mg/mL was independently prepared using a separate PS80 stock solution (solution B), also at 2 mg/mL in water, and similarly diluted with water.

Spiked samples were prepared by introducing PS80 reference material into mAb-1 solutions to achieve a consistent mAb-1 concentration of 20 mg/mL, with 5 distinct PS80 spike levels (0.05, 0.075, 0.15, 0.25, and 0.5 mg/mL). These samples were utilized to evaluate accuracy, precision, and linearity, with robustness assessed using the 0.05 and 0.5 mg/mL spiked samples. Additionally, a high protein concentration sample (160 mg/mL mAb-2 spiked with 0.15 mg/mL PS80) was employed to assess robustness under high-concentration conditions. Specificity samples included PS80 diluted with water to 0.15 mg/mL and 6 different formulation buffers (DP, A1, A2, B1, B2, C).

Protein Precipitation

The high protein concentration sample was diluted 1 : 5 using formulation buffer without PS80. For example, 200 μL of the sample was diluted to a final volume of 1000 μL with formulation buffer. The diluted sample was then mixed with acetonitrile at a 1 : 2 volume ratio (v/v), vortexed for 1 min, and incubated in a 5°C ice bath for 1 h. Following incubation, the sample was centrifuged at 10000 rpm and 5°C for 5 min, and the resulting supernatant was transferred to a clean tube. The precipitate was washed with an equal volume of acetonitrile (e.g., 2000 μL) and vortexed for 1 min. The sample was then centrifuged at 10000 rpm and 5°C for 5 min. The resulting supernatant was combined with the initial supernatant. All collected supernatants were dried using a nitrogen blower or vacuum concentrator. Finally, the dried residue was reconstituted in water to the original volume (e.g., 200 μL).

Sample Dilution

The experiment was conducted using a sample containing 160 mg/mL mAb-2 with 0.15 mg/mL PS80. The sample was diluted 1 : 5 with water, and the injection volume of the diluted sample was 100 μL, which is 5 times the volume of the undiluted sample. For the high-concentration protein sample, the PS80 injection content remained the same, but the protein concentration was reduced.

Injection Protocol and Data Analysis

Next, we injected the PS80 test control continuously until a typical and stable PS80 chromatogram was obtained. For each complete sequence, the following injections were performed: a blank (1 injection), standard solutions (5 concentrations ranging from low to high, with 1 injection for each), PS80 test control (replicate injections), samples 1 to n (2 replicates for each sample, with up to 5 samples in each sequence), and PS80 test control (2 replicates).

We used the logarithmic values of the peak area and the log value of the corresponding PS80 concentration at each concentration level to generate a linear fit standard curve. It was not necessary to constrain the standard curve to pass through the origin. The following equations were used to calculate the unknown concentration.

  

$$\begin{array}{c}\text{Log value of sample concentration}= (\text{log value of the peak}\\ \text{area of PS80}-\text{Y-axis intercept})/\text{slope of the standard curve.}\end{array}$$

  

$$\text{Sample concentration}={10}^{\text{log value of sample concentration}}.$$

RESULTS

The method was validated using 2 PS80 raw materials as reference materials (J. T. Baker and Nanjing Well). The method validation included specificity, accuracy, precision, linearity, range, and robustness; the experimental data can be found in Supplementary Tables 1, 2, and 3. Furthermore, we analyzed the overall accuracy and precision of J. T. Baker and Nanjing Well to demonstrate the differences in method validation of PS80 from different manufacturer sources. The limit of detection and limit of quantitation were not included, in accordance with International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) guideline Q2, as this is a content determination method with a defined measurement range. The criteria for methodological validation were as follows: specificity—no significant interfering peaks were observed in the area of interest in different mAb product formulations analyzed across 4 different laboratories; accuracy—spike recovery rates at all tested concentration levels were required to fall within the acceptable range of 80–120%; precision—the relative standard deviation (RSD) for replicate measurements at each spike concentration level should not exceed 10%; linearity—the calibration curve was required to demonstrate a correlation coefficient (R) ≥0.99, while spiked samples were required to maintain R ≥0.98; robustness—variations in PS80 quantification results between robustness conditions and standard conditions were required to remain within ±15% relative difference.

System Suitability

The following predefined system suitability criteria were applied: (1) no interfering peak in the blank at the area of interest (PS80 main peak), (2) the correlation coefficient of the standard curve must be ≥0.99, (3) the concentration of the PS80 test control sample should fall within the target range ±10% for each injection, (4) the ratio of low to high PS80 control injections at the beginning and end of the control set should be no less than 0.9, and (5) the ratio of the mean value of bracketing PS80 product controls (low/high) should not be less than 0.9. The mean results from the beginning and end sets can be used to calculate the bracketing result ratio. All validation steps met these requirements.

Specificity

Specificity samples included PS80 diluted with water to 0.15 mg/mL, and 6 different formulation buffers (DP, A1, A2, B1, B2, C) were analyzed across 4 different laboratories. No significant interfering peaks were observed in the area of interest (Fig. 1).

Fig. 1. Specificity Chromatograms from J. T. Baker (A) and Nanjing Well (B)

Accuracy

To assess the accuracy of the method, we performed 3 injections at each PS80 concentration (20 mg/mL mAb-1 spiked with 0.050, 0.075, 0.150, 0.250, and 0.500 mg/mL PS80) from both J. T. Baker and Nanjing Well. PS80 recovery was calculated by comparing the measured concentration to the theoretical concentration for each level. The individual recovery rates of PS80 from J. T. Baker for each of the 4 laboratories (n = 3) along with the overall recovery rate (n = 12) are presented in Fig. 2A. Similarly, the individual recovery rates of PS80 from Nanjing Well for each of the 4 laboratories (n = 3) and the overall recovery rate (n = 12) are shown in Fig. 2B. The total recovery rates of PS80 from both manufacturers (n = 6) and all recovery results (n = 24) are summarized in Fig. 2C and Table 2 (n = 24).

Fig. 2. Accuracy of Samples Spiked with PS80 from J. T. Baker and Nanjing Well

(A) J. T. Baker, (B) Nanjing Well, and (C) Total.

Table 2. Summary of Accuracy Data

Concentration (mg/mL)	0.050	0.075	0.150	0.250	0.500
Lab 1 (%)	103.0	101.1	105.7	104.7	101.4
Lab 2 (%)	109.6	104.6	103.5	106.5	101.3
Lab 3 (%)	103.5	102.6	103.4	101.7	99.4
Lab 4 (%)	111.6	101.9	100.7	99.8	100.2
Mean (%)	106.9	102.6	103.3	103.2	100.6

As shown in Fig. 2 and Table 2, all recovery rates at each PS80 concentration level from the 4 labs were within the criterion range of 80–120%, which is considered an acceptable accuracy.

Precision

Precision was evaluated using the same dataset as for accuracy. For repeatability (intra-laboratory precision) analyses of spiked samples containing PS80 sourced from either J. T. Baker or Nanjing Well, as shown in Figs. 3A–3C, the Relative Standard Deviation (RSD) was calculated for same-concentration replicates (n = 3) within individual laboratories and subsequently determined for all intra-laboratory samples (n = 6, representing 3 samples per manufacturer). Regarding intermediate precision (inter-laboratory precision) evaluations across 4 laboratories for spiked samples containing PS80 from either J. T. Baker or Nanjing Well, as shown in Figs. 3D and 3E, RSD was computed for identical-concentration measurements (n = 12, with 3 measurements per laboratory) across the 4 laboratories and analyzed for the complete dataset (n = 24, encompassing 3 measurements per laboratory from 4 laboratories and 2 manufacturers), as shown in Fig. 3F and Table 3.

Fig. 3. Precision of Samples Spiked with PS80 from J. T. Baker and Nanjing Well

(A) Repeatability of J. T. Baker, (B) repeatability of Nanjing Well, (C) repeatability of total, (D) intermediate precision of J. T. Baker, (E) intermediate precision of Nanjing Well, (F) intermediate precision of total.

Table 3. Summary of Precision Data

Concentration (mg/mL)	0.050	0.075	0.150	0.250	0.500
RSD of Lab 1 (%)	2.7	1.2	0.7	1.4	0.9
RSD of Lab 2 (%)	5.6	2.4	4.4	2.4	2.3
RSD of Lab 3 (%)	1.1	3.2	2.2	0.7	1.8
RSD of Lab 4 (%)	4.2	3.3	1.1	0.9	1.7
Intermediate precision (%)	5.1	2.8	2.9	2.9	1.8

As shown in Table 3 and Fig. 3, the repeatability and intermediate precision results passed predefined acceptance criteria (i.e., 15%).

Linearity

Linearity was evaluated using the same dataset as for accuracy. The curves for standards and spiked samples were generated using the log peak area versus log concentration formula (see “Injection protocol and data analysis”). Both the PS80 standards (J. T. Baker’s PS80 and Nanjing Well’s PS80) were assessed, and the resulting curve (standard curve) was used to calculate the measured values of PS80 in spiked samples. This also enabled plotting the curve (spiked sample linearity) with the measured values on the X-axis and the theoretical values on the Y-axis.

As shown in Tables 4–7, the linearity results of the standard curve in all 4 laboratories were greater than 0.99, which is considered acceptable. The linearity of the spiked samples also exceeded 0.99, meeting the criteria of R ≥0.98.

Table 4. Standard Curve of J. T. Baker’s Polysorbate 80 (PS80)

Standard curve	Lab 1	Lab 2	Lab 3	Lab 4
Slope	1.3405	1.6253	1.3546	1.2687
Y-intercept	2.9435	7.4236	0.0806	7.2648
R	0.9997	0.9999	0.9997	0.9993
R2	0.9994	0.9998	0.9994	0.9986

Table 5. Summary of Spiked Sample Linearity Data for J. T. Baker’s Polysorbate 80 (PS80)

Curve of spiked sample	Lab 1	Lab 2	Lab 3	Lab 4
Slope	1.3500	1.5424	1.3194	1.2224
Y-intercept	2.9193	7.3894	0.1698	7.2447
R	0.9997	0.9992	0.9999	0.9995
R2	0.9995	0.9985	0.9998	0.9991

Table 6. Standard Curve of Nanjing Well’s Polysorbate 80 (PS80)

Standard curve	Lab 1	Lab 2	Lab 3	Lab 4
Slope	1.3520	1.7090	1.3347	1.2737
Y-intercept	2.9122	7.5264	0.1304	7.2883
R	0.9998	0.9993	0.9995	0.9997
R2	0.9996	0.9986	0.9991	0.9995

Table 7. Summary of Spiked Sample Linearity Data for Nanjing Well’s Polysorbate 80 (PS80)

Curve of spiked sample	Lab 1	Lab 2	Lab 3	Lab 4
Slope	1.3456	1.7156	1.346	1.2221
Y-intercept	2.9275	7.5490	0.1228	7.2572
R	0.9997	0.9997	0.9998	0.9990
R2	0.9995	0.9994	0.9997	0.9980

Range

The summary of the cross-lab validation results demonstrates that the PS80 concentration range of 0.05–0.5 mg/mL was linear, precise, and accurate for the 20 mg/mL mAb product formulations.

Robustness

The robustness study was conducted by varying multiple critical test conditions, such as changing instrument parameters and mobile phase concentrations. For each experiment, 1 critical test condition was altered, and the changes in the target condition were compared to the robustness conditions. Each sample was injected twice, and the mean value was calculated. The relative difference in autosampler stability was calculated using Equation (1), while the relative difference for all other robustness conditions was calculated using Equation (2).

  

$$\begin{array}{l}\mathrm{R}\mathrm{e}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{a}\mathrm{u}\mathrm{t}\mathrm{o}\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{b}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}\%=\\ \frac{\left(\begin{array}{l}|\mathrm{P}\mathrm{S}80\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{T}0\mathrm{p}\mathrm{o}\mathrm{i}\mathrm{n}\mathrm{t}-\\ \mathrm{P}\mathrm{S}80\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{T}24\mathrm{p}\mathrm{o}\mathrm{i}\mathrm{n}\mathrm{t}|\end{array}\right)}{\mathrm{P}\mathrm{S}80\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{T}0\mathrm{p}\mathrm{o}\mathrm{i}\mathrm{n}\mathrm{t}}\times 100,\end{array}$$(1)

  

$$\begin{array}{l}\mathrm{R}\mathrm{e}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{o}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{r}\mathrm{o}\mathrm{b}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{n}\mathrm{e}\mathrm{s}\mathrm{s}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{s}\%=\\ \frac{\left(\begin{array}{l}|\mathrm{P}\mathrm{S}80\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}-\\ \mathrm{P}\mathrm{S}80\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{o}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{b}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{n}\mathrm{e}\mathrm{s}\mathrm{s}\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{a}\mathrm{m}\mathrm{e}\mathrm{t}\mathrm{e}\mathrm{r}|\end{array}\right)}{\mathrm{P}\mathrm{S}80\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}\times 100\end{array}$$(2)

As shown in Tables 8 and 9, the robustness conditions tested included sample stability, HPLC system parameters, mobile phase concentration, and ELSD detector parameters. Under all these conditions, the difference between the target and robustness results remained within 15%, which meets the acceptable validation acceptance criteria.

Table 8. Summary of Robustness Data for J. T. Baker’s Polysorbate 80 (PS80)

Group	Lab 1 (%)		Lab 2 (%)		Lab 3 (%)		Lab 4 (%)
PS80 concentration (mg/mL)	0.05	0.5	0.05	0.5	0.05	0.5	0.05	0.5
Autosampler stability (24 h)	4.18	1.70	6.12	9.09	6.34	4.29	2.08	1.68
Column temperature: −5°C	0.10	2.38	13.29	2.17	3.05	3.21	0.40	2.39
Column temperature: +5°C	1.45	2.04	10.12	1.91	4.29	2.72	0.00	0.06
Formic acid conc.: −10%	2.03	0.37	10.85	3.63	3.44	1.81	2.10	0.02
Formic acid conc.: +10%	0.56	1.53	2.09	2.01	11.97	5.55	1.30	0.19
Flow rate: −10%	5.72	2.31	8.64	1.31	4.87	2.50	3.10	0.77
Flow rate: +10%	4.11	1.81	7.65	0.69	3.79	3.34	4.70	0.30
ELSD EVAP: −3%	3.11	2.28	7.68	2.27	5.10	1.93	2.80	1.04
ELSD EVAP: +3%	0.06	2.52	2.05	1.48	4.75	0.57	3.80	0.34
ELSD gas flow: −10%	2.71	0.15	8.16	1.84	0.77	2.17	0.10	1.40
ELSD gas flow: +10%	1.75	1.98	3.27	1.92	2.05	0.17	3.64	2.52
ELSD NEB: −5°C	0.33	1.39	NA	NA	0.85	2.50	3.55	2.35
ELSD NEB: +5°C	1.03	2.29	NA	NA	7.27	1.20	0.68	1.51

Table 9. Summary of Robustness Data for Nanjing Well’s Polysorbate 80 (PS80)

Group	Lab 1 (%)		Lab 2 (%)		Lab 3 (%)		Lab 4 (%)
Polysorbate 80 concentration (mg/mL)	0.05	0.5	0.05	0.5	0.05	0.5	0.05	0.5
Autosampler stability (24 h)	3.00	1.32	1.85	2.87	1.57	0.86	0.98	0.62
Column temperature: −5°C	0.23	1.87	5.37	6.24	2.86	1.22	1.00	5.22
Column temperature: + 5°C	0.06	0.74	4.76	5.68	3.38	0.47	1.30	1.30
Formic acid conc.: −10%	1.80	2.72	9.19	2.14	6.71	4.26	0.30	1.99
Formic acid conc.: +10%	0.00	1.82	7.43	1.98	1.76	1.23	3.20	0.94
Flow rate: −10%	2.16	1.71	0.27	1.04	1.30	2.60	0.80	0.17
Flow rate: +10%	0.16	2.55	0.37	0.93	4.61	1.92	2.91	1.41
ELSD EVAP: −3%	0.27	3.36	7.66	1.09	2.34	4.87	5.20	2.02
ELSD EVAP: +3%	0.54	3.61	6.07	3.62	6.35	4.64	1.89	1.88
ELSD gas flow: −10%	5.67	1.48	3.57	0.29	1.41	2.08	1.83	0.08
ELSD gas flow: +10%	0.17	2.81	1.47	1.42	3.30	3.27	1.77	3.13
ELSD NEB: −5°C	0.91	1.14	NA	NA	3.10	1.60	1.39	0.31
ELSD NEB: +5°C	5.00	2.59	NA	NA	4.02	1.44	5.72	2.39

High Concentration Matrix Sample Test

High protein concentration samples (160 mg/mL mAb-2 with 0.15 mg/mL PS80) were prepared using 2 different procedures: precipitation and dilution (“Protein precipitation” and “Sample dilution”, respectively). To assess the method’s accuracy, we evaluated the recovery rate of the diluted sample, yielding a recovery rate of 93–108%. The recovery rate for the precipitated sample ranged from 83 to 106%. Both precipitation and dilution were therefore considered acceptable techniques, as the results fell within the pre-defined validation acceptance criteria of 80–120%.

Sample dilution is the recommended procedure for high-concentration matrix samples, while the precipitation procedure is considered an alternative due to the observed under-recovery, which may be caused by the more complex sample preparation process.

We repeated the optimization experiments and found that the recovery was considerably better than direct injection at nearly 100 versus 80%. However, if the sample is diluted without being made up to volume, the PS80 becomes difficult to quantify. Generally, we recommend either diluting the sample or injecting it after removing the protein if the mAb concentration exceeds 20 mg/mL.

DISCUSSION

The HPLC–ELSD method is commonly used to measure PS80 content in mAbs and is widely adopted by major pharmaceutical companies. According to the ICH guideline Q2, the measurement of PS80, as studied here, is classified as an assay (content determination). Key parameters to be verified include specificity, accuracy, precision, linearity, and range, while the limit of quantification and the limit of detection are not required. Additionally, because the same method can yield different results across laboratories, comprehensive methodological validation conducted in collaboration with multiple laboratories provides a more thorough assessment of the method’s applicability.

The method was validated for neat sample injection. However, high mAb concentrations may interfere with the PS80 measurement, potentially causing an interfering peak in the area of interest and/or poor PS80 recovery (data not shown). To mitigate this interference, we incorporated 2 additional experiments into the method robustness study: reducing the mAb concentration by dilution and performing protein precipitation using acetonitrile.

An ELSD-based method for determining PS80 concentration in mAb product formulations was successfully developed and validated. The method demonstrates good linearity, precision, and accuracy across a concentration range of 0.05–0.50 mg/mL PS80 in the presence of 20 mg/mL mAb. The robustness of the method was evaluated across different types of instruments, detectors, 2 PS80 reference materials, various mAb molecules (IgG), and 4 pharmaceutical QC labs.

The method validation data revealed slightly higher variation in the lower linearity range of 0.05 mg/mL PS80, suggesting that the protein in the sample matrix interferes with PS80, leading to over-recovery. This interference is dependent on the protein-to-PS80 ratio and increases with higher protein content and/or lower PS80 content. To overcome protein interference—especially for dosages with high protein concentrations, such as subcutaneous formulations—additional pre-treatment methods were implemented, including dilution (while ensuring the PS80 concentration remained within the linearity range) and precipitation. Both sample preparation techniques demonstrated acceptable PS80 recovery and extended the method’s usability for mAb concentrations exceeding 20 mg/mL. However, the precipitation procedure is more time- and labor-intensive, leading to slightly higher PS80 under-recovery compared to the dilution procedure. Furthermore, the available accuracy data indicate that a mAb-to-PS80 ratio ranging from 40 to 400 (0.05–0.5 mg/mL PS80 in the presence of 20 mg/mL mAb) is supported without requiring additional sample dilution.

Colorimetry is time-consuming and prone to significant systematic errors, while RPLC and GC require either protein removal or hydrolysis of PS80 into fatty acids. Therefore, compared with these methods for quantifying PS80 in therapeutic mAbs, this approach offers greater simplicity, faster analysis, and eliminates the need for complex sample pretreatment.^20–22) We have conducted method validation at 4 different sites, which significantly reinforces the validity of the findings. Additionally, the study is strengthened by the use of diverse antibodies and PS80 to validate the methodology. However, limitations exist, such as the analysis of different sample matrices and the use of real clinical samples. These experiments are currently ongoing and will be reported in future publications.

CONCLUSION

We present an HPLC method utilizing ELSD detection to quantify PS80 in mAbs. The method was independently validated across 4 different laboratories and met all ICH Q2 method validation requirements. It is a fast, reliable, and robust approach, suitable as a platform method for determining PS80 concentration in mAb products. The system is applicable to a range of IgG antibodies and various sources of PS80 materials. The linear dynamic range is from 0.05 to 0.5 mg/mL PS80, and the method remains robust for samples up to 160 mg/mL mAb. This method has considerable potential value for the pharmaceutical community, and future work will focus on evaluating samples during manufacturing quality control.

DECLARATIONS

Funding

This work was supported by the State Key Laboratory of Drug Regulatory Science Project (2025SKLDRS0318).

Author Contributions

Xiaojuan Yu and Chuanfei Yu designed the project; Xiaojuan Yu and Feiyu Wang performed the experiments; Xiaojuan Yu, Lan Wang, and Chuanfei Yu wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

REFERENCES

• 1)  Lehtimaki M, Ortega-Rodriguez U, Centola G, Rao VA. Reduced IgG1 antibody left-twisted antiparallel β-sheet structure stability occurs under metal-catalyzed oxidation conditions in the presence of polysorbates. Mol. Pharm., 22, 2623–2638 (2025).
• 2)  Honemann MN, Wendler J, Graf T, Bathke A, Bell CH. Monitoring polysorbate hydrolysis in biopharmaceuticals using a QC-ready free fatty acid quantification method. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci., 1116, 1–8 (2019).
• 3)  Zhang S, Riccardi C, Carlson D, Kamen D, Graham KS, Shameem M, Bak H, Xiao H, Li N. Prediction of long-term polysorbate degradation according to short-term degradation kinetics. MAbs, 15, 2232486 (2023).
• 4)  Desai R, Jain R, Dandekar P. Surfactants reduce aggregation of monoclonal antibodies in cell culture medium with improvement in performance of mammalian cell culture. Biotechnol. Prog., 39, e3370 (2023).
• 5)  Zürcher D, Caduff S, Aurand L, Capasso Palmiero U, Wuchner K, Arosio P. Comparison of the protective effect of polysorbates, poloxamer and brij on antibody stability against different interfaces. J. Pharm. Sci., 112, 2853–2862 (2023).
• 6)  Felix MN, Waerner T, Lakatos D, Reisinger B, Fischer S, Garidel P. Polysorbates degrading enzymes in biotherapeutics—a current status and future perspectives. Front. Bioeng. Biotechnol., 12, 1490276 (2025).
• 7)  Zürcher D, Wuchner K, Arosio P. Mitigation strategies against antibody aggregation induced by oleic acid in liquid formulations. Mol. Pharm., 21, 5761–5771 (2024).
• 8)  Hipper E, Diederichs T, Kaiser W, Lehmann F, Buske J, Hinderberger D, Garidel P. Visible light triggers the formation of reactive oxygen species in monoclonal antibody formulations. Int. J. Pharm., 661, 124392 (2024).
• 9)  Kishore RSK, Kiese S, Fischer S, Pappenberger A, Grauschopf U, Mahler H-C. The degradation of polysorbates 20 and 80 and its potential impact on the stability of biotherapeutics. Pharm. Res., 28, 1194–1210 (2011).
• 10)  Tomlinson A, Demeule B, Lin B, Yadav S. Polysorbate 20 degradation in biopharmaceutical formulations: quantification of free fatty acids, characterization of particulates, and insights into the degradation mechanism. Mol. Pharm., 12, 3805–3815 (2015).
• 11)  Mondal B, Kote M, Lunagariya C, Patel M. Development of a simple high performance liquid chromatography (HPLC)/evaporative light scattering detector (ELSD) method to determine Polysorbate 80 in a pharmaceutical formulation. Saudi Pharm. J., 28, 325–328 (2020).
• 12)  Bai L, Zhang Y, Zhang C, Lu Y, Li Z, Huang G, Meng B. Investigation of excipients impact on polysorbate 80 degradation in biopharmaceutical formulation buffers. J. Pharm. Biomed. Anal., 233, 115496 (2023).
• 13)  Doyle M, Barnes A, Larson NR, Liu H, Yi L. Development of UPLC-UV-ELSD method for fatty acid profiling in polysorbate 80 and confirmation of the presence of conjugated fatty acids by mass spectrometry, UV absorbance and proton nuclear magnetic resonance spectroscopy. J. Pharm. Sci., 112, 2393–2403 (2023).
• 14)  Zegota MM, Schuster G, De Pra M, Mullner T, Menzen T, Steiner F, Hawe A. High throughput multidimensional liquid chromatography approach for online protein removal and characterization of polysorbates and poloxamer in monoclonal antibody formulations. J. Chromatogr. A, 1720, 464777 (2024).
• 15)  Schilling K, Pawellek R, Lovejoy K, Muellner T, Holzgrabe U. Influence of charged aerosol detector instrument settings on the ultra-high-performance liquid chromatography analysis of fatty acids in polysorbate 80. J. Chromatogr. A, 1576, 58–66 (2018).
• 16)  Hewitt D, Alvarez M, Robinson K, Ji J, Wang YJ, Kao YH, Zhang T. Mixed-mode and reversed-phase liquid chromatography-tandem mass spectrometry methodologies to study composition and base hydrolysis of polysorbate 20 and 80. J. Chromatogr. A, 1218, 2138–2145 (2011).
• 17)  Magnusson LE, Risley DS, Koropchak JA. Aerosol-based detectors for liquid chromatography. J. Chromatogr. A, 1421, 68–81 (2015).
• 18)  Schilling K, Holzgrabe U. Recent applications of the charged aerosol detector for liquid chromatography in drug quality control. J. Chromatogr. A, 1619, 460911 (2020).
• 19)  Shende N, Karale A, Bhagade S, Gulhane A, Bore P, Marathe P, Chakraborty S, Mallya AD, Dhere RM. Evaluation of a sensitive GC-MS method to detect polysorbate 80 in vaccine preparation. J. Pharm. Biomed. Anal., 183, 113126 (2020).
• 20)  Hewitt D, Zhang T, Kao Y. Quantitation of polysorbate 20 in protein solutions using mixed-mode chromatography and evaporative light scattering detection. J. Chromatogr. A, 1215, 156–160 (2008).
• 21)  Paschen CA, Klemm D, Graf T, Kopf R, Pinto C, Muller C, Bell CH, Pfaff J. Simultaneous quantification of polysorbate 20 and poloxamer 188 in biopharmaceutical formulations using evaporative light scattering detection. J. Pharm. Biomed. Anal., 192, 113640 (2021).
• 22)  Zhang R, Wang Y, Tan L, Zhang HY, Yang M. Analysis of polysorbate 80 and its related compounds by RP-HPLC with ELSD and MS detection. J. Chromatogr. Sci., 50, 598–607 (2012).