Comparative Thermodynamics of the Reversible Self-Association of Therapeutic mAbs Reveal Opposing Roles for Linked Proton- and Ion-Binding Events.

PURPOSE: Reversible self-association (RSA) has long been a concern in therapeutic monoclonal antibody (mAb) development. Because RSA typically occurs at high mAb concentrations, accurate assessment of the underlying interaction parameters requires explicitly addressing hydrodynamic and thermodynamic nonideality. We previously examined the thermodynamics of RSA for two mAbs, C and E, in phosphate buffered saline (PBS). Here we continue to explore the mechanistic aspects of RSA by examining the thermodynamics of both mAbs under reduced pH and salt conditions. METHODS: Dynamic light scattering and sedimentation velocity (SV) studies were conducted for both mAbs at multiple protein concentrations and temperatures, with the SV data analyzed via global fitting to determine best-fit models, interaction energetics, and nonideality contributions. RESULTS: We find that mAb C self-associates isodesmically irrespective of temperature, and that association is enthalpically driven but entropically penalized. Conversely, mAb E self-associates cooperatively and via a monomer-dimer-tetramer-hexamer reaction pathway. Moreover, all mAb E reactions are entropically driven and enthalpically modest or minimal. CONCLUSIONS: The thermodynamics for mAb C self-association are classically seen as originating from van der Waals interactions and hydrogen bonding. However, relative to the energetics we determined in PBS, self-association must also be linked to proton release and/or ion uptake events. For mAb E, the thermodynamics implicate electrostatic interactions. Furthermore, self-association is instead linked to proton uptake and/or ion release, and primarily by tetramers and hexamers. Finally, although the origins of mAb E cooperativity remain unclear, ring formation remains a possibility whereas linear polymerization reactions can be eliminated.

Mechanistic Studies and Formulation Mitigations of Adeno-associated Virus Capsid Rupture During Freezing and Thawing: Mechanisms of Freeze/Thaw Induced AAV Rupture.

Gene therapies delivered using adeno-associated virus (AAV) vectors are showing promise for many diseases. Frozen AAV drug products are exposed to freeze-thaw (F/T) cycles during manufacturing, storage, and distribution. In this work we studied the mechanisms of AAV capsid rupture during F/T. We found that exposure to interfaces, exacerbated by F/T, and the mechanical force of excipient devitrification correlated with AAV capsid rupture during F/T. There was no impact of pH shifts, cryo-concentration, or cold-denaturation. Results were similar for AAV8 and AAV9. With these mechanistic insights we identified three formulation mitigation approaches. Addition of ≥0.0005% w/v poloxamer 188 (P188) eliminated substantial recovery losses (up to ∼60% without P188) and minimized rupture to ≤1% per F/T cycle. Elimination of exothermic devitrification events during rewarming, either by formulating with a low buffer concentration, or by adding a cryoprotectant further reduced rupture during F/T. Rupture of AAV9 was <0.2% per F/T cycle in a formulation with 1 mM phosphate, 4.4 mM dextrose, electrolytes, and 0.001% P188 at pH 7.2. Rupture of AAV8 was not detected when formulated with 4% sucrose, 100 mM salt, and 0.001% P188 at pH 7.4. These results provide insights into effective strategies for stabilizing AAVs against rupture during F/T.

Impact of Time Out of Intended Storage and Freeze-thaw Rates on the Stability of Adeno-associated Virus 8 and 9.

There are an increasing number of clinical studies evaluating different adeno-associated virus (AAV) serotypes as vectors for gene therapy. Long-term frozen storage can maximize the stability of AAV. Freeze-thaw (F/T) cycles and exposures to room temperature (RT) and refrigerated conditions occur during manufacturing, labeling, and clinical use. In this work we exposed AAV8 and AAV9 at low and high concentrations to five F/T cycles compounded with RT and refrigerated holds in a 'daisy chain' time out of intended storage (TOIS) stability study, which may be a best practice in early development. We also evaluated the impact of 5 F/T cycles for multiple permutations of fast and slow cooling and rewarming rates. The quality attributes of AAV8 and AAV9 remained within acceptable ranges after the daisy chain TOIS and F/T rate studies. Potency and concentration were unchanged within method variability. There was a minor increase in non-encapsidated ('free') DNA released from AAV8 after F/T in a phosphate-buffered saline formulation. DNA release during F/T was minimized in a formulation with a low buffer concentration and was not detected in a formulation containing sucrose. We conclude that AAV8 and AAV9 have stability profiles that are suitable for manufacturing and clinical development.

Development of a stable lyophilized adeno-associated virus gene therapy formulation.

Adeno-associated viruses (AAV) are among the most actively investigated vectors for gene therapy. Supply of early clinical studies with frozen drug product (DP) can accelerate timelines and minimize degradation risks. In the long-term, logistical challenges of frozen DP may limit patient access. In this work, we developed a lyophilized (freeze-dried) formulation of AAV. The mass concentration of AAV is typically low, and AAV also requires a minimum ionic strength to inhibit aggregation. These factors result in a low collapse temperature, which is limiting to lyophilization. Mannitol crystallization was found to cause extensive degradation and potency loss of AAV during the freezing step. With further development, we determined that AAV could be lyophilized in a sucrose and citrate formulation with a more desirable high glass transition temperature of the dried cake. An optimal residual moisture range (1-3%) was found to be critical to maintaining AAV8 stability. Glycerol was found to protect AAV8 from over-drying by preventing capsid damage and genome DNA release. A lyophilized formulation was identified that maintained potency for 24 months at 2-8 °C, indicating the feasibility of a dried formulation for AAV gene therapy.

Quantitation of Trace Levels of DNA Released from Disrupted Adeno-Associated Virus Gene Therapy Vectors.

Adeno-associated virus (AAV) vectors for gene therapy have potential to provide a durable treatment response for a number of diseases with unmet need. DNA is released from AAV capsids at high temperatures. Less is known about DNA release that may occur under conditions relevant to clinical and commercial manufacturing, storage, and distribution. In this work we developed and applied a sensitive fluorescent dye-based method to quantitate trace levels of DNA released from AAV capsids. The method was used to characterize the impact of manufacturing process steps on the increase (up to 1.5%) and removal (down to 0.2%) of free DNA. Free DNA increased by 0.3% per day at 37 °C and by 0.4% per freeze/thaw cycle in a phosphate-buffered saline formulation. When stored for 2 years at different temperatures, free DNA remained low (<0.6%) at both ≤ -60 °C and at 2-8 °C but was higher (2.6%) when the same sample was stored at -20 °C. The dye-based method may be used to further characterize release of free DNA for different processes, formulations, and stress conditions. Overall, release of free DNA was a relatively minor degradation pathway under the conditions studied in this work.

Energetic Dissection of Mab-Specific Reversible Self-Association Reveals Unique Thermodynamic Signatures.

PURPOSE: Reversible self-association (RSA) remains a challenge in the development of therapeutic monoclonal antibodies (mAbs). We recently analyzed the energetics of RSA for five IgG mAbs (designated as A-E) under matched conditions and using orthogonal methods. Here we examine the thermodynamics of RSA for two of the mAbs that showed the strongest evidence of RSA (mAbs C and E) to identify underlying mechanisms. METHODS: Concentration-dependent dynamic light scattering and sedimentation velocity (SV) studies were carried out for each mAb over a range of temperatures. Because self-association was weak, the SV data were globally analyzed via direct boundary fitting to identify best-fit models, accurately determine interaction energetics, and account for the confounding effects of thermodynamic and hydrodynamic nonideality. RESULTS: mAb C undergoes isodesmic self-association at all temperatures examined, with the energetics indicative of an enthalpically-driven reaction offset by a significant entropic penalty. By contrast, mAb E undergoes monomer-dimer self-association, with the reaction being entropically-driven and comprised of only a small enthalpic contribution. CONCLUSIONS: Classical interpretations implicate van der Waals interactions and H-bond formation for mAb C RSA, and electrostatic interactions for mAb E. However, noting that RSA is likely coupled to additional equilibria, we also discuss the limitations of such interpretations.

Analytical characterization of coformulated antibodies as combination therapy.

When two therapeutic agents are combined in a single formulation, i.e., coformulated, the quality and safety of the individual agents must be preserved. Here we describe an approach to evaluate the quality attributes of two individual monoclonal antibodies (mAbs), designated mAb-A and mAb-B, in coformulation. The mAbs were fractionated from heat-stressed coformulated drug product (DP) by hydrophobic interaction chromatography. Each purified mAb fraction was then compared with mAb-A and mAb-B in their individual formulations from the same drug substance sources used to make the coformulated DP lot, which was subjected to the same stress conditions. Product variants were evaluated and compared by using several analytical tests, including high-performance size exclusion chromatography (HPSEC), reducing and nonreducing gel electrophoresis, ion-exchange chromatography, capillary isoelectric focusing, and peptide mapping with mass spectrometry. Intermolecular interactions in coformulated and photostressed DPs were studied by evaluating aggregates fractionated from coformulated DP by HPSEC. Aggregate fractions of coformulated DP contained dimers, but not coaggregates, of mAb-A or mAb-B. Moreover, extensive assays for higher-order structure and biological interactions confirmed that there was no interaction between the two mAb molecules in the coformulation. These results demonstrate that the two coformulated therapeutic mAbs had the same quality attributes as the individually formulated mAb-A and mAb-B, no new quality attributes were formed, and no physicochemical, intermolecular, or biological interactions occurred between the two components. The approach described here can be used to monitor the product quality of other coformulated antibodies.

Monoclonal Antibody Dimers Induced by Low pH, Heat, or Light Exposure Are Not Immunogenic Upon Subcutaneous Administration in a Mouse Model.

The presence of protein aggregates is commonly believed to be an important risk factor for immunogenicity of therapeutic proteins. Among all types of aggregates, dimers are relatively abundant in most commercialized monoclonal antibody (mAb) products. The aim of this study was to investigate the immunogenicity of artificially created mAb dimers relative to that of unstressed and stressed mAb monomers. A monoclonal murine IgG1 (mIgG1) antibody was exposed to low pH, elevated temperature, or UV irradiation to induce dimerization. Dimers and monomers were purified via size-exclusion chromatography. Physicochemical analysis revealed that upon all stress conditions, new deamidation or oxidation or both of amino acids occurred. Nevertheless, the secondary and tertiary structures of all obtained dimers were similar to those of unstressed mIgG1. Isolated dimers were administered subcutaneously in Balb/c mice, and development of antidrug antibodies and accumulation of follicular T helper cells in draining lymph nodes and spleens were determined. None of the tested dimers or stressed monomers were found to be more immunogenic than the unstressed control in our mouse model. In conclusion, both dimers and monomers generated by using 3 different stress factors have a low immunogenicity similar to that of the unstressed monomers.

Investigation of cathepsin D-mAb interactions using a combined experimental and computational tool set.

Cathepsin D has been identified as a challenge to remove in downstream bioprocessing of monoclonal antibodies (mAbs) due to interactions with some mAbs. This study focused on investigating the mechanisms of interaction between cathepsin D and two industrial mAbs using a combined experimental and computational approach. Surface plasmon resonance was used to study the impact of pH and salt concentration on these protein-protein interactions. While salt had a moderate effect on the interactions with one of the mAbs, the other mAb demonstrated highly salt-dependent association behavior. Cathepsin D binding to the mAbs was also seen to be highly pH dependent, with operation at pH 9 resulting in a significant decrease in the binding affinity. Protein-protein docking simulations identified three interaction sites on both mAbs; near the complementarity determining region (CDR), in the hinge, and in the CH 3 domain. In contrast, only one face of cathepsin D was identified to interact with all the three sites on the mAbs. Surface property analysis revealed that the binding regions on the mAbs contained strong hydrophobic clusters and were predominantly negatively charged. In contrast, the binding site on cathepsin D was determined to be highly positively charged and hydrophobic, indicating that these protein-protein interactions were likely due to a combination of hydrophobic and electrostatic interactions. Finally, covalent crosslinking coupled with mass spectrometry was used to validate the docking predictions and to further investigate the regions of interaction involved in mAb-cathepsin D binding. A strong agreement was observed between the two approaches, and the CDR loops were identified to be important for cathepsin D interactions. This study establishes a combined experimental and computational platform that can be used to probe mAb-host cell protein (HCP) interactions of importance in biomanufacturing.

Submicron Size Particles of a Murine Monoclonal Antibody Are More Immunogenic Than Soluble Oligomers or Micron Size Particles Upon Subcutaneous Administration in Mice.

Protein aggregates are one of the several risk factors for undesired immunogenicity of biopharmaceuticals. However, it remains unclear which features determine whether aggregates will trigger an unwanted immune response. The aim of this study was to determine the effect of aggregates' size on their relative immunogenicity. A monoclonal murine IgG1 was stressed by exposure to low pH and elevated temperature followed by stirring to obtain aggregates widely differing in size. Aggregate fractions enriched in soluble oligomers, submicron size particles and micron size particles were isolated via centrifugation or size-exclusion chromatography and characterized physicochemically. The secondary and tertiary structures of aggregates were altered in a similar way for all the fractions, while no substantial chemical degradation was observed. Development of anti-drug antibodies was measured after subcutaneous administration of each enriched fraction to BALB/c mice. Among all tested fractions, the most immunogenic was the one highly enriched in submicron size particles (∼100-1000 nm). Fractions composed of micron size (>1-100 µm) particles or soluble oligomers (<100 nm) were not immunogenic under the dosing regimen studied in this work. These results show that aggregate size is an important factor for protein immunogenicity.

Structural insights into the mechanism of action of a biparatopic anti-HER2 antibody.

Pathways of human epidermal growth factor (EGF) receptors are activated upon ligand-dependent or -independent homo- or heterodimerization and their subsequent transphosphorylation. Overexpression of these receptors positively correlates with transphosphorylation rates and increased tumor growth rates. MEDI4276, an anti-human epidermal growth factor receptor 2 (HER2) biparatopic antibody-drug conjugate, has two paratopes within each antibody arm. One, 39S, is aiming at the HER2 site involved in receptor dimerization and the second, single chain fragment (scFv), mimicking trastuzumab. Here we present the cocrystal structure of the 39S Fab-HER2 complex and, along with biophysical and functional assays, determine the corresponding epitope of MEDI4276 and its underlying mechanism of action. Our results reveal that MEDI4276's uniqueness is based first on the ability of its 39S paratope to block HER2 homo- or heterodimerization and second on its ability to cluster the receptors on the surface of receptor-overexpressing cells.

Determination of Interaction Parameters for Reversibly Self-Associating Antibodies: A Comparative Analysis.

Monoclonal antibodies (mAbs) represent a major class of biotherapeutics and are the fastest growing category of biologic drugs on the market. However, mAb development and formulation are often impeded by reversible self-association (RSA), defined as the dynamic exchange of monomers with native-state oligomers. Here, we present a comparative analysis of the self-association properties for 5 IgG mAbs, under matched conditions and using orthogonal methods. Concentration-dependent dynamic light scattering and sedimentation velocity studies revealed that the majority of mAbs examined exhibited weak to moderate RSA. However, because these studies were carried out at mAb concentrations in the mg/mL range, we also observed significant nonideality. Noting that nonideality frequently masks RSA and vice versa, we conducted direct boundary fitting of the sedimentation velocity data to determine stoichiometric binding models, interaction affinities, and nonideality terms for each mAb. These analyses revealed equilibrium constants from micromolar to millimolar and stoichiometric models from monomer-dimer to isodesmic. Moreover, even for those mAbs described by identical models, we observed distinct kinetics of self-association. The accuracy of the models and their corresponding equilibrium constants were addressed using sedimentation equilibrium and simulations. Overall, these results serve as the starting point for the comparative dissection of RSA mechanisms in therapeutic mAbs.

Biophysical characterization of influenza A virions.

Antigenic drift of the influenza A virus requires that vaccine production is targeted to the strains circulating each year. Live-attenuated influenza A vaccine manufacturing is used to produce intact virions with the surface antigens of the circulating strains. Influenza A typically contains a large percentage (>90%) of non-infective virions. The ribonucleoprotein (RNP) content, virion structure, and aggregation are factors that are thought to have an impact on infectivity. However, these factors are difficult to study because of the intrinsic variability in virion size, shape and overall structural integrity. Negative stain TEM for total particle counts and cryoTEM for detailed size/structural analysis are established benchmark techniques for virus characterization. Other methods may be valuable for certain sample types or circumstances. The aim of this work is to establish a benchmark comparison between orthogonal biophysical techniques for particle counts, population size distribution, structural integrity, and aggregate levels. NTA and FFF-MALS rapidly provided total counts, size distribution, and aggregate/elongated virion content. CryoTEM with size analysis and fraction counting yielded detailed information about the pleomorphism of the sample. The structural integrity of virions was inferred from multi-signal AUC-SV and CryoTEM. The current work provides a comparative assessment and a baseline for the selection of biophysical tools for the determination of particle counts, aggregation and pleomorphic characteristics of influenza A virus.

Targeted Fcγ Receptor (FcγR)-mediated Clearance by a Biparatopic Bispecific Antibody.

Soluble ligands have commonly been targeted by antibody therapeutics for cancers and other diseases. Although monoclonal antibodies targeting such ligands can block their interactions with their cognate receptors, they can also significantly increase the half-life of their ligands by FcRn-mediated antibody recycling, thereby evading ligand renal clearance and requiring increasingly high antibody doses to neutralize the increasing pool of target. To overcome this issue, we generated a bispecific/biparatopic antibody (BiSAb) that targets two different epitopes on IL-6 to block IL-6-mediated signaling. The BiSAb formed large immune complexes with IL-6 that can bind Fcγ receptors on phagocytic cells and are rapidly internalized. In addition, rapid clearance of the BiSAb·IL-6 complex was observed in mice while the parental antibodies prolonged the serum half-life of IL-6. Intravital imaging of the liver in mice confirmed that the rapid clearance of these large immune complexes was associated with Fcγ receptor-dependent binding to Kupffer cells in the liver. The approach described here provides a general strategy for therapeutic antibodies with the ability to not only neutralize but also actively drive clearance of their soluble antigens.

Impact of Mutations on the Higher Order Structure and Activity of a Recombinant Uricase.

This study explores the structural and functional changes associated with a low-temperature thermal transition of 2 engineered bacterial uricase mutants. Uricase has a noncovalent homotetrameric structure, with 4 active sites located at the interface of subunits. Using differential scanning calorimetry, a low-temperature transition was identified at 42°C for mutant A and at 33°C for mutant B. This transition was stabilized by the uricase inhibitor, oxonic acid, suggesting a strong structural relationship to the active site. For mutant B, there was a reversible loss of enzymatic activity above the low-temperature transition. Spectroscopic measurements demonstrated that there was also a reversible loss of secondary and tertiary structures and an increase in surface hydrophobicity. However, the hydrophobic core environment and the tetrameric structure were not altered over the low-temperature transition suggesting that the changes occurred primarily at the surface of the enzyme. The protein became aggregation-prone at temperatures approaching the cluster of higher-temperature melting transitions at 84°C, indicating these transitions represent a global unfolding of the protein. Our findings shed light on the structural changes that affect the uricase mechanism of action and provide new insights into how enzyme therapeutic development may be approached.

Identification of an IgG CDR sequence contributing to co-purification of the host cell protease cathepsin D.

Recombinant therapeutic monoclonal antibodies (mAbs) must be purified from host cell proteins (HCPs), DNA, and other impurities present in Chinese hamster ovary (CHO) cell culture media. HCPs can potentially result in adverse clinical responses in patients and, in specific cases, have caused degradation of the final mAb product. As reported previously, residual traces of cathepsin D caused particle formation in the final product of mAb-1. The current work was focused on identification of a primary sequence in mAb-1 responsible for the binding and consequent co-purification of trace levels of CHO cathepsin D. Surface plasmon resonance (SPR) was used to detect binding between immobilized CHO cathepsin D and a panel of mAbs. Out of 13 mAbs tested, only mAb-1 and mAb-6 bound to cathepsin D. An LYY motif in the HC CDR2 was common, yet unique, to only these two mAbs. Mutation of LYY to AAA eliminated binding of mAb-1 to cathepsin D providing confirmation that this sequence motif was involved in the binding to CHO cathepsin D. Interestingly, the binding between mAb-1 and cathepsin D was weaker than that of mAb-6, which may be related to the fact that two aspartic acid residues near the LYY motif in mAb-1 are replaced with neutral serine residues in mAb-6. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 33:140-145, 2017.

Generation and Characterization of an IgG4 Monomeric Fc Platform.

The immunoglobulin Fc region is a homodimer consisted of two sets of CH2 and CH3 domains and has been exploited to generate two-arm protein fusions with high expression yields, simplified purification processes and extended serum half-life. However, attempts to generate one-arm fusion proteins with monomeric Fc, with one set of CH2 and CH3 domains, are often plagued with challenges such as weakened binding to FcRn or partial monomer formation. Here, we demonstrate the generation of a stable IgG4 Fc monomer with a unique combination of mutations at the CH3-CH3 interface using rational design combined with in vitro evolution methodologies. In addition to size-exclusion chromatography and analytical ultracentrifugation, we used multi-angle light scattering (MALS) to show that the engineered Fc monomer exhibits excellent monodispersity. Furthermore, crystal structure analysis (PDB ID: 5HVW) reveals monomeric properties supported by disrupted interactions at the CH3-CH3 interface. Monomeric Fc fusions with Fab or scFv achieved FcRn binding and serum half-life comparable to wildtype IgG. These results demonstrate that this monomeric IgG4 Fc is a promising therapeutic platform to extend the serum half-life of proteins in a monovalent format.

Fate of Multimeric Oligomers, Submicron, and Micron Size Aggregates of Monoclonal Antibodies Upon Subcutaneous Injection in Mice.

The aim of this study was to examine the fate of differently sized protein aggregates upon subcutaneous injection in mice. A murine and a human monoclonal immunoglobulin G 1 (IgG1) antibody were labeled with a fluorescent dye and subjected to stress conditions to create aggregates. Aggregates fractionated by centrifugation or gel permeation chromatography were administered subcutaneously into SKH1 mice. The biodistribution was measured by in vivo fluorescence imaging for up to 1 week post injection. At several time points, mice were sacrificed and selected organs and tissues were collected for ex vivo analysis. Part of injected aggregated IgGs persisted much longer at the injection site than unstressed controls. Aggregate fractions containing submicron (0.1-1 µm) or micron (1-100 µm) particles were retained to a similar extent. Highly fluorescent "hot-spots" were detected 24 h post injection in spleens of mice injected with submicron aggregates of murine IgG. Submicron aggregates of human IgG showed higher accumulation in draining lymph nodes 1 h post injection than unstressed controls or micron size aggregates. For both tested proteins, aggregated fractions seemed to be eliminated from circulation more rapidly than monomeric fractions. The biodistribution of monomers isolated from solutions subjected to stress conditions was similar to that of unstressed control.

Trace levels of the CHO host cell protease cathepsin D caused particle formation in a monoclonal antibody product.

Chinese hamster ovary (CHO) cells are often used to produce therapeutic monoclonal antibodies (mAbs). CHO cells express many host cell proteins (HCPs) required for their growth. Interactions of HCPs with mAbs can sometimes result in co-purification of trace levels of 'hitchhiker' HCPs during the manufacturing process. Purified mAb-1 product produced in early stages of process optimization had high HCP levels. In addition, these lots formed delayed-onset particles containing mAb-1 and its heavy chain C-terminal fragments. Studies were performed to determine the cause of the observed particle formation and to optimize the purification for improved HCP clearance. Protease activity and inhibitor stability studies confirmed that an aspartyl protease was responsible for fragmentation of mAb-1 resulting in particle formation. An affinity resin was used to selectively capture aspartyl proteases from the mAb-1 product. Mass spectrometry identified the captured aspartyl protease as CHO cathepsin D. A wash step at high pH with salt and caprylate was implemented during the protein A affinity step to disrupt the HCP-mAb interactions and improve HCP clearance. The product at the end of purification using the optimized process had very low HCP levels, did not contain detectable protease activity, and did not form particles. Spiking of CHO cathepsin D back into mAb-1 product from the optimized process confirmed that it was the cause of the particle formation. This work demonstrated that process optimization focused on removal of HCPs was successful in eliminating particle formation in the final mAb-1 product.

Gelation of a monoclonal antibody at the silicone oil-water interface and subsequent rupture of the interfacial gel results in aggregation and particle formation.

The formation of viscoelastic gels by a monoclonal antibody (mAb) at the silicone oil-water interface was studied using interfacial shear rheology. At a concentration of 50 µg/mL, the mAb formed gels in less than 1 h, and the gelation time decreased with increasing protein concentration. To probe the effects of mechanical rupture of the interfacial gel layers, a layer of silicone oil was overlaid on the surface of aqueous solutions of mAb, and the interface was ruptured periodically with a needle. Rupture of the interfacial gel resulted in formation of subvisible particles and substantial losses of mAb monomer, which were detected by microflow imaging and quantified by size-exclusion chromatography, respectively. Resonance mass measurement showed that levels of both protein particles and silicone oil droplets increased as the gel was repeatedly ruptured with a needle. In contrast, in samples wherein the interfacial gels were not ruptured, much lower levels of aggregates and particles were detected. Addition of nonionic surfactants (polysorbate 20 or polysorbate 80) protected against aggregation and protein particle formation, with increased protection seen with increasing surfactant levels, and with the greatest inhibition observed in samples containing polysorbate 80.