Reduced IgG1 Antibody Left-Twisted Antiparallel β-Sheet Structure Stability Occurs under Metal-Catalyzed Oxidation Conditions in the Presence of Polysorbates

Abstract

Polysorbates are common surfactants in monoclonal antibody (mAb) drug products. While polysorbates assist in stabilizing and refolding proteins, oxidative stress conditions can reduce protein stability wherein polysorbate binds to the oxidized and unfolded protein. We investigated the effects of polysorbates on the higher-order structural stability of mAbs under oxidative conditions that may occur during manufacturing, storage, and use. Secondary and tertiary structures of trastuzumab and rituximab products were investigated under two oxidative conditions: metal-catalyzed oxidation (MCO; CuSO₄ and ascorbic acid) and 2,2'-azobis (2-aminidinopropane) dihydrochloride (AAPH) using either polysorbate-containing formulations or after polysorbate depletion. Higher-order structures were predicted from the collected circular dichroism spectra with an algorithm optimized for β-sheet structural predictions. Secondary structure analyses using circular dichroism at increasing temperatures demonstrated that MCO and AAPH triggered differing β-sheet structure degradation patterns. Rituximab products were more sensitive to MCO compared with trastuzumab products as shown by left-twisted antiparallel β-sheet structure loss and increase in unstructured elements at lower temperatures. AAPH-exposed drugs tended to have distinct unfolding states compared with the MCO-treated drugs as shown by the increase in parallel β-sheet structures for AAPH and decreased parallel β-sheet structures with MCO. Polysorbate depletion transiently improved the stability of MCO-treated material as shown by delayed circular dichroism (CD) signal degradation at 202 nm and improved peak area of the antibody monomer by nonreducing capillary electrophoresis sodium dodecyl sulfate (nrCE-SDS) and peak intensity of intact antibody in matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) analysis. The improved stability of detergent-depleted material is traced to improved stability in the predicted left-twisted β-sheet structural elements. Our data further highlights the need for formulation studies that consider the impact of polysorbate binding and/or degradation for specific drug products under stress conditions such as metal-catalyzed oxidation.

Keywords: higher-order structure; monoclonal antibody; oxidation; polysorbate; stability.

Figure 1.

Depletion of polysorbate from trastuzumab product (A) and rituximab product (F) samples using DiI assays is shown using paired students t-test to calculate significance with α level of 0.05 (n=5). Spectral 202 nm peak signal of MCO and AAPH-treated original formulation and polysorbate-depleted formulation trastuzumab (B, C), trastuzumab-pkbr (D, E), rituximab (G, H), and rituximab-abbs (I, J) samples. Statistical analyses with 2-way analysis of variance (ANOVA) and Dunnett’s multiple comparisons. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001 (N=3).

Figure 2.

Trastuzumab and trastuzumab-pkbr left-twisted antiparallel β-sheet structure content stability is reduced by MCO and AAPH treatment similarly in the original formulation and polysorbate-depleted samples as the temperature is increased. The secondary structure predictions for β-sheet structures are analyzed separately for original formulation and polysorbate-depleted material using 2-way ANOVA and Dunnett’s multiple comparisons. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001 (N=3).

Figure 3.

Rituximab and rituximab-abbs left-twisted antiparallel β-sheet structure content is reduced at 20 and 35°C in the original formulation. The secondary structure predictions for β-sheet structures are analyzed separately for original formulation and polysorbate depleted material using 2-way ANOVA and Dunnett’s multiple comparisons. * P<0.05, ** P<0.01, *** P<0.001, and **** P<0.0001 (N=3).

Figure 4.

Trastuzumab and trastuzumab-pkbr parallel β-sheet structure content is variable and tends to increase in AAPH-treated samples with the original formulation. The secondary structure predictions for β-sheet structures are analyzed separately for original formulation and polysorbate depleted material using 2-way ANOVA and Dunnett’s multiple comparisons. * P<0.05, ** P<0.01, *** P<0.001, and **** P<0.0001 (N=3).

Figure 5.

Rituximab and rituximab-abbs parallel β-sheet structure content is variable and tends to increase in AAPH-treated samples with the original formulation. Polysorbate-depleted formulation tends to have increased parallel β-sheet content in MCO samples. The secondary structure predictions for β-sheet structures are analyzed separately for original formulation and polysorbate-depleted material using 2-way ANOVA and Dunnett’s multiple comparisons. * P<0.05, ** P<0.01, *** P<0.001, and **** P<0.0001 (N=3).

Figure 6.

High-molecular-weight species (HMW), the main peak, and low-molecular-weight (LMW) species % peak area of trastuzumab (A), trastuzumab-pkbr (B), rituximab (C), and rituximab-abbs (D) fluorescence signal spectra before and after polysorbate depletion measured using SEC-UPLC. Original formulation and polysorbate-depleted formulation peak areas are compared using 2-way ANOVA and Dunnett’s multiple comparisons. * P<0.05, ** P<0.01, *** P<0.001, and **** P<0.0001 (N=3).

Figure 7.

Microflow imaging shows increased 1–150 μm total particle concentration due to MCO treatment, especially in rituximab and rituximab-abbs samples. The nonparametric Mann-Whitney tests show no significant differences between the original formulation and polysorbate-depleted material (N=3).

Figure 8.

Electropherogram of nonreducing CE-SDS of trastuzumab (A) and trastuzumab-pkbr (E) overlaying the control with MCO- and AAPH-treated material in original formulation and polysorbate-depleted material. Quantified percent main, low molecular weight (LMW), and high molecular weight (HMW) peak areas of nonreduced CE-SDS electropherograms for trastuzumab (B-D) and trastuzumab-pkbr (F-H), with and without detergent and by MCO and AAPH degradation at 3 and 24 h. Statistical analysis with 2-way ANOVA and Sidak’s multiple comparisons. (P<0.05=*, P<0.01=**, P<0.001=***, and P<0.0001=****, N=3).

Figure 9.

Electropherogram of nonreducing CE-SDS of rituximab (A) and rituximab-abbs (E) overlaying the control with MCO- and AAPH-treated material in original formulation and polysorbate-depleted material. #=internal control. Quantified percent main, low molecular weight (LMW), and high molecular weight (HMW) peak areas of nonreduced CE-SDS electropherograms for rituximab (B-D), and rituximab-abbs (F-H) with and without detergent and by MCO degradation at 3 and 24 h and by AAPH degradation at 3 hours and 24 hours. Statistical analysis with 2-way ANOVA and Sidak’s multiple comparisons. (P<0.05=*, P<0.01=**, P<0.001=***, and P<0.0001=****, N=3).