Physicochemical Comparison of Kolliphor HS 15, ELP, and Conventional Surfactants for Antibody Stabilization in Biopharmaceutical Formulations

Abstract

Stabilization of therapeutic protein formulations with nonionic surfactants such as Polysorbate 20 (PS20), Polysorbate 80 (PS80), or Poloxamer 188 (P188) is imperative to avoid critical loss of the active pharmaceutical ingredient by aggregation or adsorption onto different types of interfaces. In the present work, we have characterized the interfacial activity of the aforementioned surfactants, alone and in competition with antibodies, in comparison to two other excipients with approval for use in parenteral applications, Kolliphor HS 15 (HS15) and Kolliphor ELP (ELP). To this end, we applied a comprehensive suite of experimental techniques, including tensiometry, interfacial rheology, and quartz-crystal microbalance with dissipation monitoring (QCM-D). The obtained data shows important differences between the surfactants as well as a clear influence of the type of interface considered on the observed behavior. In order to link these physicochemical results to the performance of the chosen surfactants in the stabilization of antibodies, we performed another series of tests to quantify protein aggregation (i.e., the formation of (sub)visible particles in formulations under stress) as well as the release of oil from siliconized vials. In addition, the stability of the surfactants against enzymatic degradation was investigated. It is demonstrated that HS15 can compete with the widely used polysorbates in terms of interfacial activity and protein stabilization, while offering higher robustness against degradation by a lipase and an esterase. On the other hand, P188 shows poor interfacial activity but can still suppress the aggregation of at least some proteins, indicating that different mechanisms of stabilization are at play. Our findings and the broad spectrum of tests described in this work are instructive toward a better understanding of protein stabilization in distinct primary packaging systems through surfactants in aqueous formulations.

Keywords: antibodies; container closure; quartz-crystal microbalance; surface activity; surfactants.

1

Plot of the surface tension as a function of the surfactant concentration in buffered aqueous solution (20 mM histidine chloride, pH 6.0) at 23 °C in the absence of antibodies.

2

Characterization of the competitive activity of surfactants and polyclonal IgG antibodies at the air/water interface by pendant drop tensiometry and dilatational rheology at 23 °C. (a) Surface tension measured for 0.1% surfactant solutions in histidine chloride buffer (20 mM, pH 6.0) in the absence (white bars) and presence (gray bars) of 0.1% IgG. (b) Phase shift determined for the same samples by interfacial rheology at 0.1 Hz. The dashed horizontal lines represent values obtained for 0.1% solutions of IgG in buffer without added surfactants. Data are given as averaged results of at least two independent measurements with corresponding standard deviations. Individual values were determined after an initial equilibration period of 600 s.

3

Characterization of the competitive activity of surfactants and polyclonal IgG antibodies at the silicone oil/water interface by pendant drop tensiometry and dilatational rheology at 23 °C. (a) Interfacial tension measured for 0.1% surfactant solutions in histidine chloride buffer (20 mM, pH 6.0) in the absence (white bars) and presence (gray bars) of 0.1% IgG. (b) Phase shift determined for the same samples by interfacial rheology at 0.1 Hz. The dashed horizontal lines represent values obtained for 0.1% solutions of IgG in buffer without added surfactants. Data are given as averaged results of at least two independent measurements with corresponding standard deviations. Individual values were determined after an initial equilibration period of 600 s.

4

Adsorption of surfactants and IgG antibodies on borosilicate glass (a–c) and PDMS (d–f) surfaces at 23 °C, as traced by QCM-D experiments. (a, d): Time-dependent changes in apparent mass at the solid/liquid interfaces during incubation with 0.1% (w/v) solutions of different surfactants in histidine chloride buffer (20 mM, pH 6.0) for 60 min, followed by rinsing with neat buffer for another 60 min. Dashed vertical lines signify the times at which the liquid medium was switched from buffer to surfactant solution and back to buffer, respectively. (b, e): Time-dependent changes in apparent mass during incubation of the interfaces obtained at the end of the surfactant adsorption/desorption processes shown in (a, c) with 0.1% (w/v) solutions of bovine IgG in histidine chloride buffer (20 mM, pH 6.0) for 30 min and subsequent rinse with neat buffer for another 30 min (where the vertical lines indicate the respective times of switching between liquid media). Note that frequency (and thus mass) changes caused by prior surfactant adsorption were deducted from the signal during subsequent protein adsorption to obtain common initial baselines. (c, f): Bar plots show the amounts of surfactants and antibodies adsorbed on borosilicate glass and PDMS surfaces after incubation and subsequent rinse. Horizontal lines represent the amounts of IgG bound to blank substrates without prior surfactant adsorption after incubation (dashed) and rinse (dotted). All apparent mass changes were calculated using the fifth overtone of the fundamental resonance frequency of the QCM-D sensors via Sauerbrey’s equation. Data are given as averaged results of at least two independently performed measurements with corresponding standard deviations.

5

Quantification of subvisible particles formed in aqueous formulations of bovine IgG after horizontal shaking for 15 h in glass vials in the presence of PS20, PS80, P188, HS15, and ELP at two different concentrations by means of flow-imaging microscopy. For comparison, the results obtained in control experiments without the added surfactants are also shown. Bars represent average values of particle counts for different size regimes (as indicated), with corresponding standard deviations derived from analyses of four independently prepared vials per formulation, which were each measured in duplicate after 1:10 dilution with buffer.

6

Quantification of subvisible particles with sizes ≥2 μm formed in aqueous formulations of (a) panitumumab and (b) abatacept under quiescent conditions (no shaking; after incubation for 4 d) as well as after vertical shaking in glass vials with (after incubation for 2 d) and without (after incubation for 4 d) added glass beads in the presence of PS20, PS80, P188 and HS15 at two different concentrations by means of flow-imaging microscopy. For comparison, the results obtained in control experiments without added surfactants, as well as the corresponding particle concentration detected at the beginning of each experiment (T0) are also included. Bars represent the results of single measurements for each tested formulation. Note that all surfactants were tested at a common concentration of 0.04% (w/v) as well as at a second concentration meant to represent their typical dosage in pharmaceutical formulations (i.e., much higher for P188 and lower for all other surfactants). ,

7

Quantification of particle formation and oil release from siliconized glass vials in aqueous formulations of bovine IgG after horizontal shaking for 15 h in the presence of PS20, PS80, P188, HS15 and ELP at a surfactant concentration of 0.05% by means of flow-imaging microscopy. Bars and dots represent the number densities of all detected particles (2–300 μm, including silicone oil droplets; left y-axis in linear scale) and silicone oil droplets only (5–300 μm; right y-axis in log scale), respectively, that were counted per vial. All values are given as averages with corresponding standard deviations derived from the analyses of four independently prepared vials per formulation, which were each measured in duplicate. Particles and droplets could be distinguished through a filter tree based on differences in the contrast and shape of the detected species. For comparison, the results obtained in control experiments without added surfactants are also included.

8

Results of forced degradation of PS20, PS80 and HS15 by different enzymes (AOL, PPL, and PLE), expressed as peak areas of mono- and polyesters relative to the surfactant reference sample [%] after incubation for 24 h in the presence of (a) low and (b) high enzyme concentrations. Bars represent the results of single measurements for each tested enzyme concentration. Note that 100% means no degradation, while complete degradation would correspond to 0%.