Subvisible Particles Research Articles

Existence of a superior polysorbate fraction in respect to protein stabilization and particle formation?

Biologicals including monoclonal antibodies are the current flagships in pharmaceutical industry. However, they are exposed to a multitude of destabilization conditions like for instance hydrophobic interfaces, leading to reduced biological activity. Polysorbates are commonly applied to effectively stabilize these active pharmaceutical ingredients against colloidal stress. Nevertheless, chemical instability of polysorbate via hydrolysis or oxidation results in degradation products that might form particles via phase separation. Polysorbates are mixtures of hundreds of individual components, and recently purer quality grades with reduced variations in the fatty acid composition are available. As the protective function of polysorbate itself is not completely understood, even less is known about its individual components, raising the question of the existence of a superior polysorbate species in respect to protein stabilization or degradation susceptibility. Here, we evaluated the protective function of four main fractions of polysorbate 20 (PS20) in agitation studies with monoclonal antibodies, followed by particle analysis as well as protein and polysorbate content determination. The commercially-available inherent mixtures PS20 high purity and PS20 all-laurate, as well as the fraction isosorbide-POE-monolaurate showed superior protection against mechanical–induced stress (visual inspection and turbidity) at the air–water interface in comparison to sole sorbitan-POE-monolaurate, -dilaurate, and -trilaurate. Fractions composed mainly of higher–order esters like sorbitan-POE-dilaurate and sorbitan-POE-trilaurate indicated high turbidities as indication for subvisible and small particles accompanied by a reduced protein monomer content after agitation. For the isosorbide-POE-monolaurates as well as for the inherent polysorbate mixtures no obvious differences in protein content and protein aggregation (SEC) were observed, reflecting the observations from visual appearance. However, absolute polysorbate concentrations vary drastically between different species in the actual formulations. As there are still open questions in respect to protein specificity or regarding mixtures versus individual components of PS20, further studies must be performed, to gain a better understanding of a “generalized” stabilizing effect of polysorbates on monoclonal antibodies. The knowledge of the characteristics of individual polysorbate species can have the potential to pave the way to superior detergents in respect to protein stabilization and/or degradation susceptibility.

An Intra-Company Analysis of Inherent Particles in Biologicals Shapes the Protein Particle Mitigation Strategy Across Development Stages

To better understand protein aggregation and inherent particle formation in the biologics pipeline at Novartis, a cross-functional team collected and analyzed historical protein particle issues. Inherent particle occurrences from the past 10 years were systematically captured in a protein particle database. Where the root cause was identified, a number of product attributes (such as development stage, process step, or protein format) were trended. Several key themes were revealed: 1) there was a higher propensity for inherent particle formation with non-mAbs than with mAbs; 2) the majority of particles were detected following manufacturing at scale, and were not predicted by the small-scale studies; 3) most issues were related to visible particles, followed by subvisible particles; 4) 50% of the issues were manufacturing related. These learnings became the foundation of a particle mitigation strategy across development and technical transfer, and resulted in a set of preventive actions. Overall, this study provides further insight into a recognized industry challenge and hopes to inspire the biopharmaceutical industry to transparently share their experiences with inherent particles formation.

Evaluation of subvisible particles in human immunoglobulin and lipid nanoparticles repackaged from a multi-dose vial using plastic syringes

The multi-dose vial (MDV) is widely used for most biopharmaceuticals that are repackaged in plastic syringes before use. However, subvisible particle formation with the use of plastic syringes containing silicone oil (SO syringes) for handling therapeutic proteins can be problematic. This study aimed to evaluate the extent of and trends in microparticle (>1 μm) formation and accumulation in repackaged syringes from MDVs containing human immunoglobulin (IgG) and lipid nanoparticles (LNPs). Light obscuration (LO) and flow imaging (FI) were used to analyze the microparticles. The number of microparticles observed with the use SO syringes was greater than that with SO-free syringes, and the number of microparticles continuously increased as did the number of times of repackaging in syringes for both drugs. However, a large variation was observed across different brands of SO syringes. In contrast, using a different technique of drug withdrawal from the vial significantly reduced the number of microparticles. Furthermore, the use of filter-integrated needles or the inclusion of stabilizers such as acetyl-arginine and Tween 20 into the formulation also helped reduce particle formation.

Micro-flow imaging multi-instrument evaluation for sub-visible particle detection

Sub-visible particles (SVPs) in pharmaceutical products are a critical quality attribute, and therefore should be monitored during development. Although light obscuration (LO) and microscopic particle count tests are the primary pharmacopeial methods used to quantify SVPs, flow imaging methods like Micro-Flow Imaging (MFI™) appear to overcome shortcomings of LO such as limited sensitivity concerning smaller translucent SVPs in the size range < 10 µm. Nowadays, MFI™ is routinely utilized during development of biologicals. Oftentimes multiple devices are distributed across several laboratories and departments. This poses challenges in data interpretation and consistency as well as in the use of multiple devices for one purpose. In this study, we systematically evaluated seven MFI™ instruments concerning their counting and size precision and accuracy, using an inter-comparable approach to mimic daily working routine. Therefore, we investigated three different types of particles (i) NIST certified counting standards, (ii) protein-coated particles, and (iii) stress-induced particles from a monoclonal antibody. We compared the results to alternative particle detection methods: LO and Backgrounded Membrane Imaging (BMI). Our results showed that the precision and accuracy of particle count and size, as well as the comparability of instruments, depended on the particle source and its material properties. The various MFI™ instruments investigated showed high precision (<15 %) and data generated on different instruments were of the same order of magnitude within pharmacopeial relevant size ranges for NIST certified counting standards. However, we found limitations in the upper and lower detection limits, contrary to the limits claimed by the manufacturer. In addition, proteinaceous and protein-containing particles showed statistically significant differences in particle counts, while the measured particle diameters of all sizes were quite consistent.

Sensitivity and Uncertainty Analysis of Micro-Flow Imaging for Sub-Visible Particle Measurements Using Artificial Neural Network.

During biopharmaceutical drug manufacturing, storage, and distribution, proteins in both liquid and solid dosage forms go through various processes that could lead to protein aggregation. The extent of aggregation in the sub-micron range can be measured by analyzing a liquid or post-reconstituted powder sample using Micro-Flow Imaging (MFI) technique. MFI is widely used in biopharmaceutical industries due to its high sensitivity in detecting and analyzing particle size distribution. However, the MFI's sensitivity to various factors makes accurate measurement challenging. Therefore, in light of the inherent variability of the method, this work aims to explore the capabilities of an adopted coupled sensitivity analysis and machine learning algorithm to quantify the influencing factors on the formed sub-visible particles and method variability. The proposed algorithm consists of two interconnected components, namely a surrogate model with a neural network and a sensitivity analyzer. A machine learning tool based on artificial neural networks (ANN) is constructed with MFI data. The best fit with an optimized configuration is found. Sensitivity and uncertainty analysis is performed using this network as the surrogate model to understand the impacts of input parameters on MFI data. Results reveal the most impactful reconstitution preparation factors and others that are masked by the instrument variabilities. It is shown that instrument inaccuracy is a function of size category, with higher variabilities associated with larger size ranges. Utilizing this tool while assessing the sensitivity of outputs to various parameters, measurement variabilities for analytical characterization tests can be quantified.

Interfacial Effects during Phase Change in Multiple Levitated Tetrahydrofuran Hydrate Droplets

Recent strategies developed to examine the nucleation of crystal structures like tetrahydrofuran (THF) hydrates without the effects of a solid interface have included acoustic levitation, where only a liquid-gas interface initially exists. However, the ability now exists to levitate and freeze multiple droplets simultaneously, which could reveal interdroplet effects and provide further insight into interfacial nucleation phenomena. In this study, using direct digital and infrared imaging techniques, the freezing of up to three simultaneous THF hydrate droplets was investigated for the first time. Nucleation was initiated at the aqueous solution-air interface. Two pseudo-heterogeneous mechanisms created additional nucleation interfaces: one from cavitation effects entraining microbubbles and another from subvisible ice particles, also called hydrate-nucleating particles (HNPs), impacting the droplet surface. For systems containing droplets in both the second and third positions, nucleation was statistically simultaneous between all droplets. This effect may have been caused by the high liquid-solid interfacial pressures that developed at nucleation, causing some cracking in the initial hydrate shell around the droplet and releasing additional HNPs (now of hydrate) into the air. During crystallization, the THF hydrate droplets developed a completely white opacity, termed optical clarity loss (OCL). It was suggested that high hydrate growth rates within the droplet resulted in the capture of tiny air bubbles within the solid phase. In turn, light refraction through many smaller bubbles resulted in the OCL. These bubbles created structural inhomogeneities, which may explain how the volumetric expansion of the droplets upon complete solidification was 23.6% compared with 7.4% in pure, stationary THF hydrate systems. Finally, the thermal gradient that developed between the top and bottom of the droplet during melting resulted in a surface tension gradient along the air-liquid interface. In turn, convective cells developed within the droplet, causing it to spin rapidly about the horizontal axis.

Physicochemical Stability of Generic Thiotepa Concentrate and Ready-to-Administer Infusion Solutions for Conditioning Treatment

The objective of this study was to determine the physicochemical in-use stability of recently approved Thiotepa Riemser concentrate in the original vial and diluted ready-to-administer (RTA) infusion solutions in prefilled glucose 5% and 0.9% NaCl polyolefin bags. Thiotepa Riemser 10 mg/mL concentrates and infusion solutions (1 mg/mL, 2 mg/mL, 3 mg/mL) were prepared in triplicate and stored at 2-8 °C or 25 °C for 14 days. Thiotepa concentrations were determined using a stability-indicating RP-HPLC assay. In parallel, pH and osmolality were measured. Sub-visible particles were counted on day 0 and 14. Thiotepa Riemser concentrate was revealed to be stable for 14 days when stored at 2-8 °C, or for 24 h when stored at 25 °C. Thiotepa concentrations in infusion solutions stored at 2-8 °C remained above 95% of the initial concentrations for at least 14 days, regardless of the type of vehicle solution. When stored at 25 °C, thiotepa infusion solutions in glucose 5% proved to be physicochemically stable for 3 days (1 mg/mL), 5 days (2 mg/mL) or 7 days (3 mg/mL). Thiotepa infusion solutions in 0.9% NaCl remained physicochemically stable for 5 days (1 mg/mL) or 7 days (2 mg/mL, 3 mg/mL). At these points in time, the specification limit of ≤0.6% monochloro-adduct was fulfilled. In parallel, an elevation of the pH values was registered. Thiotepa concentrates and infusion solutions should be stored at 2-8 °C due to temperature-dependent physicochemical stability, and for microbiological reasons. Glucose 5% infusion solution is recommended as a diluent, and stability-improving nominal 2 mg/mL to 3 mg/mL thiotepa concentrations should be obtained.

Impact of Stress on the Immunogenic Potential of Adalimumab

Monoclonal antibodies against tumor necrosis factor-alpha (TNFα) are widely used for treatment of inflammatory diseases. However, despite the inhibitory effect this class of drugs has on the immune system, anti-drug antibodies are often formed with continuous use. Particles formed during stress conditions, which can be used to simulate storage and handling conditions of commercial antibodies, have previously been associated with the formation of anti-drug antibodies. This study investigates the relationship between particles, oligomerization, folding and chemical degradation on the in vitro cytokine response toward the TNFα inhibitor adalimumab. Adalimumab aggregates generated using stir and heat stress were fractionated into distinct sub-populations, and their structure and immunogenic potential were evaluated. A chemically degraded sample of adalimumab was included to compare particle composition with the milder accelerated heat and stir stressed conditions. Particles from stressed adalimumab samples induced elevated cytokine levels and CD4+ T cell proliferation in vitro compared to non-stressed samples. Samples enriched with both submicron and subvisible particles of adalimumab induced the strongest cytokine release and the strongest CD4+ T cell proliferation despite maintaining some TNFα inhibitory functionality. Samples that were stressed and subsequently purified of subvisible and submicron particles did not elicit a significantly higher cytokine response or show increased CD4+ T cell proliferation compared to a non-stressed sample. Oxidation-induced chemical modifications in adalimumab, mainly in Met, His, Trp, and Tyr, were not found to be sufficient in absence of particle formation to induce increased CD4+ T cell proliferation or cytokine release despite less decreased TNFα inhibitory activity of adalimumab. These observations provide further evidence that particles do indeed potentiate the immunogenic potential of adalimumab.

Physicochemical and Biological Stability Assessment of SB11 (Ranibizumab Biosimilar) Under Ambient and In-Use Storage for Intravitreal Administration.

SB11 (Byooviz™) is a ranibizumab biosimilar that acts as a vascular endothelial growth factor (VEGF)-A inhibitor. Stability data for unopened SB11 vials at room temperature are limited and no data are available for SB11 withdrawn into syringes (in-use) for intravitreal administration. SB11 stability was assessed in two different settings: unopened vials stored at 30 ± 2°C/65 ± 5% relative humidity (RH) for 2months, and in-use SB11 withdrawn into syringes stored at 5 ± 3°C for 98days and then 25 ± 2°C/60 ± 5% RH for 24h. The product was stored in the absence of light, and the experimental design followed International Conference on Harmonization and European Medicines Agency requirements for stability evaluation of biological products. Analysis included visual appearance (color, clarity, and presence of visible particles), pH, protein concentration (A280) and purity (size-exclusion high-pressure liquid chromatography, capillary electrophoresis-sodium dodecyl sulfate, imaged capillary isoelectric focusing), biological activity (VEGF binding and neutralization), and safety (sub-visible particulates). Except for charge variants in unopened vials at room temperature after 1month by US standards, all results met the stability acceptance criteria (US and EU) for both unopened vials and for in-use SB11. There were no major changes in terms of physicochemical stability, biological activity and sub-visible particulates. SB11 was stable for longer periods and at higher temperatures than what is stated in the labels of the reference product (Lucentis) and SB11. The physicochemical properties, biological activity, and sub-visible particulates of SB11 in both tested settings (unopened vials at room temperature and in-use product withdrawn into syringes) were maintained under the described storage periods. This information can help to avoid unnecessary delays in patient treatment without any loss in quality and biological activity, lower the workload of health care providers and reduce costs associated with drug waste.

Structural, kinetic, and thermodynamic aspects of insulin aggregation.

Given the significance of protein aggregation in proteinopathies and the development of therapeutic protein pharmaceuticals, revamped interest in assessing and modelling the aggregation kinetics has been observed. Quantitative analysis of aggregation includes data of gradual monomeric depletion followed by the formation of subvisible particles. Kinetic and thermodynamic studies are essential to gain key insights into the aggregation process. Despite being the medical marvel in the world of diabetes, insulin suffers from the challenge of aggregation. Physicochemical stresses are experienced by insulin during industrial formulation, storage, delivery, and transport, considerably impacting product quality, efficacy, and effectiveness. The present review briefly describes the pathways, mathematical kinetic models, and thermodynamics of protein misfolding and aggregation. With a specific focus on insulin, further discussions include the structural heterogeneity and modifications of the intermediates incurred during insulin fibrillation. Finally, different model equations to fit the kinetic data of insulin fibrillation are discussed. We believe that this review will shed light on the conditions that induce structural changes in insulin during the lag phase of fibrillation and will motivate scientists to devise strategies to block the initialization of the aggregation cascade. Subsequent abrogation of insulin fibrillation during bioprocessing will ensure stable and globally accessible insulin for efficient management of diabetes.

The Strengths of Total Holographic Video Microscopy in Detecting Sub-Visible Protein Particles in Biopharmaceuticals: A Comparison to Flow Imaging and Resonant Mass Measurement

Determination of subvisible particle (SVP) content in biopharmaceuticals is a prerequisite to ensure the quality of liquid biopharmaceutical products. Here, we present a comparison of the recently introduced holographic video microscopy (total holographic characterization, THC) with two orthogonal and well-established analytical technologies: micro flow imaging (MFI) and resonant mass measurement (RMM). The capabilities of the THC were investigated under conditions commonly applied in drug product development. Three different antibody products were used at different concentrations and formulations to cover a wide range of realistic use-cases. The comparison was particularly focused on protein aggregates to investigate the applicability of THC to this critical class of particles in drug product development. Protein concentrations up to 100 mg/ml were investigated covering a broad range of viscosity and refractive indices, both important parameters in particle detection. The comparison reveals that THC is highly sensitive to detect protein aggregates in a size range from 0.5 µm to 10 µm. THC shows a significant superiority to FI and RMM in detecting heterogenous protein aggregates which often appear as transparent and porous particles. Additionally, THC needs very small sample amount of about 30 µl and short measurement times, making it applicable for early development stages and high-throughput approaches. These results show that THC is a valuable supplement to the existing particle characterization method portfolio in drug product development.

New insight in algal cell adhesion and cake layer evolution in algal-related membrane processes: Size-fractioned particles, initial foulant seeds and EDEM simulation

A clear understanding of algal cell adhesion and cake layer evolution in algal-related membrane processes (ARMPs) is urgently required to mitigate the membrane fouling. In this study, the effect of microparticles (10 μm–30 μm), subvisible particles (0.45 μm–10 μm), and ultrafine particles (50 kDa-0.45 μm) on the membrane fouling were explored based on the filtration performance through Hermia models, thermodynamic analysis, and simulation of extended discrete element method (EDEM). The results illustrated that microparticles played an important role in algal cell aggregation and the formation of initial clusters. Intermediate blocking fouling occurred when filtrating the subvisible particle, which facilitated internal adhesion and enhanced biofilm formation. In addition, the interfacial attractive force for the initial algal adhesion was obviously increased when the membrane surfaces were in high concentration of protein and polysaccharide. Moreover, the EDEM simulation demonstrated that subsequent particles, particularly the particles with small sizes, preferred to occupy the spaces among the previously deposited particles. This study provided new insights into the contributions of size-fractioned particles to initial fouling and their influence on the successive adhesion of other contaminants.

In vitro Stability of Biosimilar Insulin Aspart SAR341402 in the Medtronic MiniMed Insulin Pumps

SAR341402 (Insulin aspart Sanofi®) is an insulin aspart biosimilar that can be used for continuous subcutaneous insulin infusion (CSII) in pump systems. The physicochemical stability of SAR341402 for CSII use was evaluated in several in vitro experiments. Insulin aspart products (SAR341402, NovoLog®, NovoRapid®) were filled into pump reservoirs and pumped through Medtronic insulin pumps (MiniMedTM 530G-Model 751, Medtronic, Northridge, CA) and their related infusion sets under simulated stress conditions, including elevated temperature and mechanical agitation on a continuously vibrating platform, up to 13 days. Samples pumped through the infusion sets and retained in reservoirs (non-pumped) were analyzed using suitable analytical methods. All products showed stable insulin aspart content and no unwanted impurities. Minor pH changes were seen in all products but were not considered relevant. A time-dependent increase in high-molecular-weight proteins and largest other insulin aspart impurities was observed for each product but each remained within acceptance limits. Concentrations of phenol and metacresol decreased but remained at levels to ensure preservative efficacy. Samples collected from the infusion sets were clear of visible particles and showed comparable subvisible particle counts. No occlusion events were observed. Leachable profiles from pump and reservoir samples were similar in all product batches. Like NovoLog®/NovoRapid®, SAR341402 demonstrates appropriate physicochemical stability when used in these insulin pump systems.

Convolutional Neural Networks Enable Highly Accurate and Automated Subvisible Particulate Classification of Biopharmaceuticals.

Quantification of subvisible particles, which are generally defined as those ranging in size from 2 to 100µm, is important as critical characteristics for biopharmaceutical formulation development. Micro Flow Imaging (MFI) provides quantifiable morphological parameters to study both the size and type of subvisible particles, including proteinaceous particles as well as non-proteinaceous features incl. silicone oil droplets, air bubble droplets, etc., thus enabling quantitative and categorical particle attribute reporting for quality control. However, limitations in routine MFI image analysis can hinder accurate subvisible particle classification. In this work, we custom-built a subvisible particle-aware Convolutional Neural Network, SVNet, which has a very small computational footprint, and achieves comparable performance to prior state-of-art image classification models. SVNet significantly improves upon current standard operating procedures for subvisible particulate assessments as confirmed by thorough real-world validation studies.

Alternative Excipients for Protein Stabilization in Protein Therapeutics: Overcoming the Limitations of Polysorbates.

Given their safety and efficiency in protecting protein integrity, polysorbates (PSs) have been the most widely used excipients for the stabilization of protein therapeutics for years. In recent decades, however, there have been numerous reports about visible or sub-visible particles in PS-containing biotherapeutic products, which is a major quality concern for parenteral drugs. Alternative excipients that are safe for parenteral administration, efficient in protecting different protein drugs against various stress conditions, effective in protein stabilization in high-concentrated liquid formulations, stable under the storage conditions for the duration of the product's shelf-life, and compatible with other formulation components and the primary packaging are highly sought after. The aim of this paper is to review potential alternative excipients from different families, including surfactants, carbohydrate- and amino acid-based excipients, synthetic amphiphilic polymers, and ionic liquids that enable protein stabilization. For each category, important characteristics such as the ability to stabilize proteins against thermal and mechanical stresses, current knowledge related to the safety profile for parenteral administration, potential interactions with other formulation components, and primary packaging are debated. Based on the provided information and the detailed discussion thereof, this paper may pave the way for the identification or development of efficient excipients for biotherapeutic protein stabilization.

Formulation mitigations for particle formation induced by enzymatic hydrolysis of polysorbate 20 in protein-based drug products: insights from a full-factorial longitudinal study

Hydrolytic degradation of the polysorbate 20 (PS20) surfactant in protein-based liquid formulations releases free fatty acids (FFAs), which can accumulate to form particles in drug products during real-time (long-term) storage. To identify formulation conditions that mitigate the risk of particle formation, we conducted a longitudinal study using purified recombinant monoclonal antibody (mAb) formulated in 24 conditions. In this real-time stability study at 5 °C, three key formulation parameters—mAb concentration, initial PS20 concentration, and pH—were varied across representative ranges in a full-factorial design. A longitudinal regression analysis was used to evaluate the effects of these parameters and their interactions on PS20 degradation (via measurements of PS20, FFAs, and PS20 ester distribution) and on particle formation (via visible particle observations and subvisible particle counts). The time-dependent onset of visible particles trended with the rise in subvisible particle counts and FFA levels and fall in PS20 concentration. In the ranges studied here, lower mAb concentration and higher initial PS20 concentration delayed the onset of particles, whereas pH had a negligible effect. These observations were consistent with the general trends predicted by our previously published FFA solubility model. Taken together, these findings highlight the complex relationships between formulation parameters, PS20 degradation, and particle formation.

Insights into the Stabilization of Interferon Alpha by Two Surfactants Revealed by STD-NMR Spectroscopy

Surfactants are commonly used in biopharmaceutical formulations to stabilize proteins against aggregation. However, the choice of a suitable surfactant for a particular protein is decided mostly empirically, and their mechanism of action on molecular level is largely unknown. Here we show that a straightforward label-free method, saturation transfer difference (STD) nuclear magnetic resonance (NMR) spectroscopy, can be used to detect protein-surfactant interactions in formulations of a model protein, interferon alpha. We find that polysorbate 20 binds with its fatty acid to interferon, and that the binding is stronger at pH closer to the isoelectric point of the protein. In contrast, we did not detect interactions between poloxamer 407 and interferon alpha. Neither of the two surfactants affected the tertiary structure and the thermal stability of the protein as evident from circular dichroism and nanoDSF measurements. Interestingly, both surfactants inhibited the formation of subvisible particles during long-term storage, but only polysorbate 20 reduced the amount of small soluble aggregates detected by size-exclusion chromatography. This proof-of-principle study demonstrates how STD-NMR can be employed to quickly assess surfactant-protein interactions and support the choice of surfactant in protein formulation.

Evaluation of a Self-Supervised Machine Learning Method for Screening of Particulate Samples: A Case Study in Liquid Formulations

Imaging is commonly used as a characterization method in the pharmaceuticals industry, including for quantifying subvisible particles in solid and liquid formulations. Extracting information beyond particle size, such as classifying morphological subpopulations, requires some type of image analysis method. Suggested methods to classify particles have been based on pre-determined morphological features or use supervised training of convolutional neural networks to learn image representations in relation to ground truth labels. Complications arising from highly complex morphologies, unforeseen classes, and time-consuming preparation of ground truth labels, are some of the challenges faced by these methods. In this work, we evaluate the application of a self-supervised contrastive learning method in studying particle images from therapeutic solutions. Unlike with supervised training, this approach does not require ground truth labels and representations are learned by comparing particle images and their augmentations. This method provides a fast and easily implementable tool of coarse screening for morphological attribute assessment. Furthermore, our analysis shows that in cases with relatively balanced datasets, a small subset of an image dataset is sufficient to train a convolutional neural network encoder capable of extracting useful image representations. It is also demonstrated that particle classes typically observed in protein solutions administered by pre-filled syringes emerge as separated clusters in the encoder's embedding space, facilitating performing tasks such as training weakly-supervised classifiers or identifying the presence of new subpopulations.

Machine learning approaches to root cause analysis, characterization, and monitoring of subvisible particles in monoclonal antibody formulations.

Processing stresses on therapeutic proteins may cause formation of subvisible particles. Different stress mechanisms generate particle populations with characteristic morphological "fingerprints,"and machine learning techniques like convolutional neural networks (CNNs) allow classification of microscopy images of these particles according to known stresses at their root cause. Using CNNs to classify novel particle types not included during network training may lead to inaccurate classification, however, using CNNs to monitor the presence of particulate matter not explicitly used in training could serve as a useful process analytical technology. We used CNNs to classify and identify the root cause of particles generated by subjecting three monoclonal antibodies (mAbs) to various common manufacturing stresses. We probed the generality of particles generated by stressing different mAbs in different formulations and showed that CNN analyses were sensitive not only to the applied stress, but also the buffer conditions and the particular mAb that generated particle populations. Thus, models trained on images of particles created with one mAb and buffer system may not provide accurate root cause analysis when applied to particles generated by other mAb and buffer systems. A lever-rule analysis of CNN-derived fingerprints was used to characterize the composition of mixtures of particle types. Finally, we monitored the temporal evolution of CNN-derived fingerprints when novel populations of particles, which were not included during training, were generated by pumping mAb solutions through a peristaltic pump.

Polysorbates versus Hydroxypropyl Beta-Cyclodextrin (HPβCD): Comparative Study on Excipient Stability and Stabilization Benefits on Monoclonal Antibodies.

Polysorbates (PS 20 and PS 80) are the most widely used surfactants in biopharmaceutical formulations to protect proteins from denaturation, aggregation, and surface adsorption. To date, around 70% of marketed therapeutic antibodies contain either PS 20 or PS 80 in their formulations. However, polysorbates are chemically diverse mixtures, which are prone to degradation by oxidation and hydrolysis to produce peroxides and fatty acids, which, in turn, induce protein oxidation, aggregation, and insoluble particle formation. These will negatively impact protein quality and stability. Thus, polysorbate degradation has emerged as one of the major challenges in the development and commercialization of therapeutic protein products. KLEPTOSE® HPβCD (hydroxypropyl beta-cyclodextrin), a new multifunctional excipient, has been shown to provide protein stabilization functions in biopharmaceutical downstream processes and in their final formulations. This study aims to evaluate HPβCD, a new molecule of its class, against polysorbates as a stabilizer in biologics formulations. In this study, the chemical stability of KLEPTOSE® HPβCDs is compared with polysorbates (20 and 80) under various stress conditions. When subjected to heat stress, HPβCDs show little change in product recovery (90.7–100.7% recovery for different HPβCDs), while polysorbates 20 and 80 show significant degradation, with only 11.5% and 7.3% undegraded product remaining, respectively. When subjected to other chemical stressors, namely, autoclave, light, and oxidative stresses, HPβCD remains almost stable, while polysorbates show more severe degradation, with 95.5% to 98.8% remaining for polysorbate 20 and 85.5% to 97.4% remaining for polysorbate 80. Further, profiling characterization and degradation analysis reveal that chemical structures of HPβCDs remain intact, while polysorbates undergo significant hydrolytic degradation and oxidation. Lastly, the physicochemical stability of monoclonal antibodies in formulations is investigated. When subjected to light stress, adalimumab, as a model mAb, formulated in the presence of HPβCD, shows a significant decrease in protein aggregation, and superior monomer and total protein recovery compared to PS 80-containing formulations. HPβCD also reduces both agitation and thermal stress-induced protein aggregation and prevents subvisible particle formation compared to PS 80.