A four-stage process that identified a hydroxypropyl beta cyclodextrin and polysorbate containing formulation that eliminated aggregation of recombinant human NELL-1 after exposure to extrinsic stresses

Vaccine formulations may contain visible and/or subvisible particles, which can vary in both size and morphology. Extrinsic particles, which are particles not part of the product such as foreign contaminants, are generally considered undesirable and should be eliminated or controlled in injectable products. However, biological products, in particular vaccines, may also contain particles that are inherent to the product. Here we focus on the characterization of visible and subvisible particles in a live, replication-deficient viral vaccine candidate against HSV genital herpes in an early developmental stage. HSV-2 viral vaccine was characterized using a panel of analytical methods, including Fourier transform infrared spectroscopy (FTIR), sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), Western blot, liquid chromatography-mass spectrometry (LC-MS), light microscopy, transmission electron microscopy (TEM), micro-flow imaging (MFI), dynamic light scattering (DLS), right angle light scattering (RALS), and intrinsic fluorescence. Particles in HSV-2 vaccine typically ranged from hundreds of nanometers to hundreds of micrometers in size and were determined to be inherent to the product. The infectious titer did not correlate with any trend in subvisible particle concentration and size distribution as shown by DLS, MFI, and TEM under stressed conditions. This suggested that particle changes in the submicron range were related to HSV-2 virion structure and had direct impact on biological activity. It was also observed that subvisible and visible particles could induce aggregation in the viral product. The temperature induced aggregation was observed by RALS, intrinsic fluorescence, and DLS. The increase of subvisible particle size with temperature could be fitted to a two-step thermokinetic model. Visible and subvisible particles were found to be inherent to the HSV-2 viral vaccine product. The mechanism of protein aggregation was discussed and a two-step thermokinetic aggregation profile was proposed. The approaches reported in this study may be applied to a variety of vaccines and other biological products, as a way to assess the consistency of the manufacturing process and identify key product quality attributes.

Isothermal chemical denaturation as a complementary tool to overcome limitations of thermal differential scanning fluorimetry in predicting physical stability of protein formulations

The subvisible and visible particles present in a solution are often classified based on size, and are quantified by the actual number of particles present rather than by weight or molar amounts. The analysis of these particles in protein therapeutics are governed by compendial methods and the regulatory agencies, and the methods available to measure them originally evolved focusing on potential safety issues, including capillary occlusion and immunogenicity, that might arise from their presence. Ultracentrifugation, size exclusion chromatography, etc., discussed in previous articles, can be used to analyze aggregates of less than 0.10 microns. This article will focus on methods for analyzing and quantitating sub visible particles (SbVP) of 2 microns or larger. At the present time there is no routine method for quantitating sub visible particles (SbVP) between 0.1 microns and 2 microns. The most common technique for quantitating the amount of subvisible particles between 2 and 100 microns is the light obscuration method. This technique can determine size and amount of particles, but cannot differentiate between the types of particles, such as protein particles, foreign material, micro bubbles or silicone oil droplets, that can be present in protein solutions. The difficulties in adapting this method, originally developed for small molecule drugs for IV administration, to protein therapeutics delivered subcutaneously is discussed. The flow imaging techniques can determine morphology and optical characteristics of the particles, but still not identify the chemical composition. Other methods that can also be used, but are applicable for characterization purposes only, are discussed. The primary method for quantitating visible particles is visual inspection, a method that can be subjective and relies on adequate training of the human inspectors. Automated methods for visible particle determination are being developed. Identification of the chemical composition of isolated particles greater than about 50 microns is possible using several micro-spectroscopic methods, and these will also be discussed.

Degradation of the non-ionic surfactant polysorbate (PS) has been reported to lead to the formation of visible and subvisible particles that affect product quality. Chemical degradation pathways of PS, such as oxidation and acid/base hydrolysis, have been previously studied, but enzymatic degradation of PS remains poorly understood. In this study, enzyme-mediated hydrolysis of the major components in a heterogeneous mixture of PS20 or PS80 was monitored using an evaporative light scattering detection-high-performance liquid chromatography method. Carboxylester hydrolases from a broad range of organisms were tested, including enzymes from Pseudomonas cepacia, Thermomyces lanuginosus, Candida antarctica, rabbit liver, and pig pancreas. Time course experiments suggested that PS hydrolysis was dependent on the order of esters (mono-, di-, or triester), the identity of the hydrophilic head group (sorbitan or isosorbide), the identity of the fatty acid ester tail (C12 vs C18:1), and the identity of the enzyme. Importantly, no PS component was completely resistant to all the carboxylester hydrolases tested here. Our results illustrate a potential fingerprint approach that could be useful in verifying or identifying enzyme-mediated PS degradation in drug substance and provide an improved understanding of the complexity of PS degradation in the presence of enzymes.

Polysorbate 20 (PS20) is a commonly used surfactant in biopharmaceutical formulations. It is a heterogeneous surfactant containing a distribution of fatty acid esters, which are subject to hydrolytic degradation, generating free fatty acids (FFAs). The FFAs can form visible or subvisible particles in drug product on stability. A previous FFA solubility model, developed by our group, predicts solubility limits for the three most prevalent FFA degradation products of PS20: lauric, myristic, and palmitic acid. The model takes into account two formulation parameters, pH and PS20 concentration, and their effect on FFA solubility. This work identifies a third parameter that has an impact on FFA solubility: PS20 ester distribution. When PS20 is hydrolytically degraded, the ester distribution of the remaining surfactant changes on stability. Ester distribution is known to influence the critical micelle concentration (CMC) of PS20 such that the monoesters have a much higher CMC compared to the higher-order esters (HOE). We hypothesize that as PS20 degrades, the CMC changes, affecting the proportion of PS20 that is present in micelles and capable of sequestering and solubilizing FFAs in these micelles. Here, PS20 was separated into monoester, HOE, and polyol fractions. The monoester and HOE fractions were mixed together to generate the mock degradation profiles of hydrolytically degraded PS20. FFA solubility was measured as a function of the concentration of these mock-degraded (MD) PS20s. The results indicate that ester distribution does have an impact on FFA solubility, especially at higher MD PS20 concentrations. HOEs solubilize up to 30 μg/mL more lauric acid than an equivalent amount of monoesters at a MD PS20 level of 0.06% w/v. With the addition of % HOE peak area fraction as a third parameter representing the ester distribution of PS20, the refined FFA solubility model more accurately predicts FFA solubility in protein formulations at 5 °C. The refined model suggests that drug products containing trace levels of host cell proteins (HCPs) that preferentially degrade HOEs of PS20 are at a higher risk of particle formation.

There is an increased interest from industry, academia and regulators for protein aggregates, subvisible and visible particles due to possible biological consequences, such as immunogenicity, altered bioactivity and modified pharmacokinetic profiles. Aggregates, subvisible and visible particles are important product instabilities, which might be present in every formulation of biotherapeutic products like monoclonal antibody solutions. Especially, the presence of subvisible particles in biotherapeutic products is currently a hot topic and it constantly gains more importance. The ultimate goal of this thesis is to develop tools and techniques in order to be able to characterize well-defined size fractions of proteinaceous subvisible particles with various desired oxidation profiles using in vivo transgenic mouse model. Up to now only a few articles were published and the available data from in vitro and in vivo experiments on aggregates and subvisible particles is often conflicting and fragmented, which impedes the development of sound conclusions. Moreover, complex mixtures of monomers, aggregates, particles and other degradants were used to draw conclusions. Only estimated values of protein particle density were used up to now in published studies although it is required to know the density of the measured particles in order to accurately calculate their dimensions and mass. The first aim of this thesis was therefore to develop a method to measure experimentally the protein particle density without extrapolation (Chapter 1). The density for commercially available standard beads (polystyrene, polymethacrylate and melamine) and a large bench of stressed proteinaceous samples was determined with the use of the resonant mass measurement instrument (RMM, Archimedes) and its ability to measure the buoyant mass of individual particles. Various fluids with increasing densities were implemented in order to determine the neutral buoyant mass where the particle density equals the fluid density. Chapter 2 reports the development of a process to isolate well-defined subvisible fractions using differential centrifugation for a model IgG1 antibody. The process to separate four fractions in the submicron and micrometer size range was developed and successfully optimized through the use of a design of experiments. The centrifugation technique was compared to an already published fractionation method using a preparative fluorescence-activated cell sorter. Efficiency, advantages and drawbacks for both methods were compared and discussed (Chapter 2). Oxidation profile of aggregates and subvisible particles seem to be an important attribute regarding induced biological consequences. That is why the next goal was to develop a method for selective oxidation of methionine and tryptophan residues in a model mAb in order to be able to delineate the effects and the contribution of individual protein modifications in the primary structure. This included a large set of experiments where different reaction conditions such as temperature of incubation, reaction time, type and concentration of oxidant (t-BHP, H2O2, AAPH) were evaluated in presence (or not) of a large excess of anti-oxidant (free amino acids) in order to protect the corresponding amino acid in a model antibody of the IgG1 subtype (Chapter 3). To complete the work, unfractionated materials and well-defined size fractions (with well-established oxidation profile) were prepared using the established tools (Chapter 1-3). Those samples were deeply characterized and injected subcutaneously into wild type and transgenic mice for immunization. Anti-drug antibody levels were measured following ELISA in order to assess the immunogenic potential of those preparations (Chapter 4).

Dissolution of Polysorbate 20 Degradation Related Free Fatty Acid Particles in Intravenous Bag Solutions.

The time-course and extent of visible particle (VP) and sub-visible particle (SVP) formation was monitored as a function of interfacial area (IA) for a model bioconjugate. To facilitate particle formation, the bioconjugate was agitated in a glass vial and exposed to IAs up to 478mm2. Since vials had equal fill and headspace volumes, the area of the air-water interface was varied by placing vials on angled blocks at 0°, 30°, 60°, or 90° from the horizontal. A significant increase in visible and sub-visible particle formation was observed with increasing air-water IA. Exposure to IAs below ∼305mm2 resulted in the formation of very few particles, while IAs > ∼305mm2 resulted in substantial particle formation. Visible and sub-visible particle morphology varied with interfacial area and time. The sub-visible particles initially increased with time but did not reach steady state; instead the initial increase was followed by complete depletion. These phenomena indicate that visible particle formation likely increased at the expense of the sub-visible particle population and demonstrate a potential link between the two particle populations for this model bioconjugate. Initiation of particle formation did not result in corresponding decreases in protein concentration or increases in soluble aggregates. However, extended agitation time resulted in a significant decrease in protein concentration.

Degradation of Polysorbate 20 by Sialate O-Acetylesterase in Monoclonal Antibody Formulations

Visible particles must be monitored as part of the control strategy for parenteral biopharmaceutical drug products. In these products, formation of protein particles is a natural occurrence. All protein drugs contain particles that vary greatly in visibility and size from invisible (sub-micron) to visible (millimeter), and pharmaceutical companies are required to monitor and minimize the presence of visible and sub-visible particles in their products. There is an industry-wide unmet need for particle standards for visual inspection of protein drugs. A new, semi-quantitative method using particle standards for assessing the levels of small, naturally occurring visible particles is presented. This method can be used during drug development to identify a formulation that minimizes particle formation and also during testing of final clinical or commercial drug product to monitor and control naturally occurring proteinaceous visible particles.

Thermal shift assay (TSA) with altered temperature has been the most widely used method for monitoring protein stability for drug research. However, there is a pressing need for isothermal techniques as alternatives. This urgent demand arises from the limitations of TSA, which can sometimes provide misleading ranking of protein stability and fail to accurately reflect protein stability under physiological conditions. Although differential scanning fluorimetry has significantly improved throughput in comparison to differential scanning calorimetry and differential static light scattering throughput, all these methods exhibit moderate sensitivity. In contrast, current isothermal chemical denaturation (ICD) techniques may not offer the same throughput capabilities as TSA, but it provides more precise information about protein stability and interactions. Unfortunately, ICD also suffers from limited sensitivity, typically in micromolar range. We have developed a novel method to overcome these challenges, namely throughput and sensitivity. The novel Förster Resonance Energy Transfer (FRET)-Probe as an external probe is highly applicable to isothermal protein stability monitoring but also to conventional TSA. We have investigated ICD for multiple proteins with focus on KRASG12C with covalent inhibitors and three chemical denaturants performed at nanomolar protein concentration. Data showed corresponding inhibitor-induced stabilization of KRASG12C to those reported by nucleotide exchange assay.

Formulations That Suppress Aggregation During Long-Term Storage of a Bispecific Antibody are Characterized by High Refoldability and Colloidal Stability

Background: Baxter received reports of visible precipitate, identified as calcium carbonate, forming during hemofiltration with Accusol 35 solution. Aim: To evaluate the potential for acute cardiopulmonary adverse effects of Accusol 35 containing exaggerated calcium carbonate particles. Methods: Anesthetized dogs underwent continuous veno-venous hemofiltration (CVVH) with Accusol 35 containing visible and subvisible particles (≥10 µm) 36 times higher than the maximum concentration specified in the European Pharmacopoeia (P-Accusol), or Accusol 35 conforming to specification (Accusol). Select cardiovascular and blood gas parameters were evaluated during CVVH. Lung tissue samples were collected following CVVH. Results: No differences were observed in cardiovascular and blood gas parameters or lung histology between P-Accusol and Accusol. Conclusion: Accusol 35 containing visible and subvisible particles (≥10 µm) 36 times higher than the maximum concentration specified in the European Pharmacopoeia resulted in no acute cardiopulmonary adverse effects compared with Accusol 35 containing no visible particles and subvisible particles within European Pharmacopoeia specification.

Screening techniques for monitoring the sub-visible particle formation of free fatty acids in biopharmaceuticals

A Newly Identified Impurity in Polysorbate 80, the Long-Chain Ketone 12-Tricosanone, Forms Visible Particles in a Biopharmaceutical Drug Product