"High-risk" host cell proteins (HCPs): A multi-company collaborative view.

Host cell proteins (HCPs) are process-related impurities that may copurify with biopharmaceutical drug products. Within this class of impurities there are some that are more problematic. These problematic HCPs can be considered high-risk and can include those that are immunogenic, biologically active, or enzymatically active with the potential to degrade either product molecules or excipients used in formulation. Some have been shown to be difficult to remove by purification. Why should the biopharmaceutical industry worry about these high-risk HCPs? What approach could be taken to understand the origin of its copurification and address these high-risk HCPs? To answer these questions, the BioPhorum Development Group HCP Workstream initiated a collaboration among its 26-company team with the goal of industry alignment around high-risk HCPs. The information gathered through literature searches, company experiences, and surveys were used to compile a list of frequently seen problematic/high-risk HCPs. These high-risk HCPs were further classified based on their potential impact into different risk categories. A step-by-step recommendation is provided for establishing a comprehensive control strategy based on risk assessments for monitoring and/or eliminating the known impurity from the process that would be beneficial to the biopharmaceutical industry.

Ranking the susceptibility of disulfide bonds in human IgG1 antibodies by reduction, differential alkylation, and LC-MS analysis.

One of the basic structural features of human IgG1 is the arrangement of the disulfide bond structure, 4 inter chain disulfide bonds in the hinge region and 12 intra chain disulfide bonds associated with twelve individual domains. Disulfide bond structure is critical for the structure, stability, and biological functions of IgG molecules. It has been known that inter chain disulfide bonds are more susceptible to reduction than intra chain disulfide bonds. However, a complete ranking of the susceptibility of disulfide bonds in IgG1 molecules is lacking. A method including reduction, differential alkylation, and liquid chromatography-mass spectrometry (LC-MS) analysis was developed and employed to investigate the complete ranking order of the susceptibility of disulfide bonds in two recombinant monoclonal antibodies. The results confirmed that inter chain disulfide bonds were more susceptible than intra chain disulfide bonds. In addition, it was observed that the disulfide bonds between the light chain and heavy chain were more susceptible than disulfide bonds between the two heavy chains. The upper disulfide bond of the two inter heavy chain disulfide bonds was more susceptible than the lower one. Furthermore, disulfide bonds in the CH2 domain were the most susceptible to reduction. Disulfide bonds in VL, CL, VH, and CH1 domains had similar and moderate susceptibility, while disulfide bonds in the CH3 domain were the least susceptible to reduction. Interestingly, a difference between IgG1kappa and IgG1lambda was also observed. The difference in the susceptibility of inter light heavy chain disulfide bonds and inter heavy chain disulfide bonds was smaller in IgG1kappa than in IgG1lambda. The intra chain disulfide bonds in the Fab region of IgG1kappa were also less susceptible than disulfide bonds in the Fab region of IgG1lambda.

Domain-level stability of an antibody monitored by reduction, differential alkylation, and mass spectrometry analysis.

Human immunoglobulin G1 (IgG1) contains 12 domains, and each has an intrachain disulfide bond that connects the two layers of antiparallel beta-sheets. These intrachain disulfide bonds are shielded from solvents under native conditions. Therefore, accessibility of the disulfide bonds to reduction under conditions that unfold antibody has the potential to be a good indicator of the thermodynamic stability of each domain. The stability of a recombinant monoclonal antibody at the domain level was investigated using a novel method involving reduction of the disulfide bonds in the presence of increasing amounts of guanidine hydrochloride and alkylation with [(12)C]iodoacetic acid, which was followed by reduction of the remaining disulfide bonds and alkylation with [(13)C]iodoacetic acid. The percentage of modification by [(12)C]iodoacetic acid of each cysteine residue was calculated using mass spectra of the cysteine-containing tryptic peptides and used to follow the unfolding of each domain. It demonstrated that the CH2 domain was the least stable domain of the antibody, whereas the CH3 domain was the most stable domain of the antibody. Other domains showed intermediate resistance to the denaturant concentration, similar to the overall unfolding transition monitored by the intrinsic tryptophan fluorescence wavelength shift.

Analysis of reduced monoclonal antibodies using size exclusion chromatography coupled with mass spectrometry.

Size-exclusion chromatography (SEC) has been widely used to detect antibody aggregates, monomer, and fragments. SEC coupled to mass spectrometry has been reported to measure the molecular weights of antibody; antibody conjugates, and antibody light chain and heavy chain. In this study, separation of antibody light chain and heavy chain by SEC and direct coupling to a mass spectrometer was further studied. It was determined that employing mobile phases containing acetonitrile, trifluoroacetic acid, and formic acid allowed the separation of antibody light chain and heavy chain after reduction by SEC. In addition, this mobile phase allowed the coupling of SEC to a mass spectrometer to obtain a direct molecular weight measurement. The application of the SEC-MS method was demonstrated by the separation of the light chain and the heavy chain of multiple recombinant monoclonal antibodies. In addition, separation of a thioether linked light chain and heavy chain from the free light chain and the free heavy chain of a recombinant monoclonal antibody after reduction was also achieved. This optimized method provided a separation of antibody light chain and heavy chain based on size and allowed a direct measurement of molecular weights by mass spectrometry. In addition, this method may help to identify peaks eluting from SEC column directly.

Identification and localization of unpaired cysteine residues in monoclonal antibodies by fluorescence labeling and mass spectrometry.

Each human IgG1 antibody contains a total of thirty-two cysteine residues. In theory, all of them are involved in disulfide bonds, and no free sulfhydryl should be detected. However, literature has suggested that the presence of low levels of free sulfhydryl groups is likely a common feature of recombinant and wild type IgG1 antibodies. Currently, there is little information correlating the presence of free sulfhydryl to specific cysteine residues. The presence of free sulfhydryl groups in five recombinant monoclonal antibodies and their locations were further investigated in the current study. Free sulfhydryl groups were first modified using 5-idoacetamidofluorescein (5-IAF), which was followed by reduction of disulfide bonds and alkylation with iodoacetic acid (IAA). This procedure allowed differentiation of free cysteine residues from cysteine residues that were involved in disulfide bonding. In addition, it allowed a sensitive fluorescence detection of peptides with free sulfhydryl groups. The locations of the free sulfhydryl groups were determined using mass spectrometry after fraction collection. The results indicated that different antibodies had different levels of free sulfhydryl due to multiple unpaired cysteine residues commonly in the constant domains. Furthermore, free sulfhydryl due to unpaired cysteine residues in the variable domains varied for different antibodies. Interestingly, free sulfhydryl was rarely associated with cysteine residues that were involved in interchain disulfide bonds.

Effect of the conserved oligosaccharides of recombinant monoclonal antibodies on the separation by protein A and protein G chromatography.

Glycosylation of the conserved asparagine residue in CH2 domains of IgG molecules is an important post-translational modification. The presence of oligosaccharides is critical for structure, stability and biological function of IgG antibodies. Effect of the glycosylation states of recombinant monoclonal antibodies on protein A and protein G chromatography was evaluated. Antibodies lacking oligosaccharides eluted later from protein A and earlier from protein G columns than antibodies with oligosaccharides using a gradient of decreasing pH. Interestingly, different types of oligosaccharides also affected the elution of the antibodies. Antibodies with high mannose type oligosaccharides were enriched in later eluting fractions from protein A and earlier eluting fractions from protein G. While antibodies with more mature oligosaccharides, such as core fucosylated biantennary complex oligosaccharides with zero (Gal 0), one (Gal 1) or two (Gal 2) terminal galactoses, were enriched in earlier eluting fractions from protein A and in the later eluting fractions from protein G. However, analysis by enzyme-linked immunosorbent assay (ELISA) revealed that antibody binding affinity to protein A and protein G was not affected by the absence or presence of oligosaccharides. It was thus concluded that the elution difference of antibodies with or without oligosaccharides and antibodies with different types of oligosaccharides were due to differential structural changes around the CH2-CH3 domain interface under the low pH conditions used for protein A and protein G chromatography.

Mass spectrometry analysis of photo-induced methionine oxidation of a recombinant human monoclonal antibody.

Oxidation of methionine (Met) residues of a recombinant fully human monoclonal antibody after exposure to light was investigated and compared with chemically induced oxidation using tert-butyl-hydroperoxide (tBHP). Met256 and Met432 in the Fc region in the samples exposed to light or incubated with tBHP were oxidized. The Fc mass spectra of the antibody exposed to light showed mainly peaks with a molecular weight (MW) increase of 32 Da, however the sample treated with tBHP showed peaks with increase of only 16 Da. These results suggested that either oxidation of one Met residue (either Met256 or Met432) catalyzed the oxidation of the second Met residue on the same heavy chain (HC) or Met residues of one HC were preferentially oxidized when the antibody was exposed to light, while Met256 and Met432 were randomly oxidized when the antibody was incubated with tBHP.

Method to differentiate asn deamidation that occurred prior to and during sample preparation of a monoclonal antibody.

Asparagine (Asn) deamidation is a major source of antibody instability and micro heterogeneity. For this reason, it is critical to accurately characterize both the levels and the sites of Asn deamidation in therapeutic antibodies. Asn deamidation is normally quantified by analyzing antibodies at the peptide level by liquid chromatography-mass spectrometry. This requires denaturation, reduction, alkylation, and enzyme digestion of the antibody prior to analysis. These steps in sample preparation may directly contribute to the total levels of Asn deamidation detected. Therefore, to obtain accurate levels and sites of Asn deamidation, it is important to determine if any deamidation occurred during the sample preparation steps. However, this could be challenging because deamidation that occurred prior to and during sample preparation resulted in peptides with the same retention times and the same molecular weight increase of 1 Da. Sample preparation was carried out in (18)O-water in the current study to differentiate between the two events of Asn deamidation. Using this method, deamidation that occurred during sample preparation resulted in a molecular weight increase of 3 Da instead of 1 Da. This molecular weight difference was readily detected by inspection of the isotopic peak cluster of the peptides containing the deamidation products, isoAsp and Asp residues. It enabled discrimination of deamidation that was due to analytical artifacts and thus determination of the level of deamidation that was present in the samples.

Glutamine deamidation of a recombinant monoclonal antibody.

Deamidation of glutamine (Gln) proceeds at a much slower rate than deamidation of asparagine (Asn) residues at peptide level. However, deamidation of Gln residues in native proteins may occur faster because of the impact of protein structure and thus plays a significant role in affecting protein stability. Gln deamidation of a recombinant monoclonal IgG1 antibody was investigated in the current study. Deamidation was determined by a molecular weight increase of 1 Da, a retention time shift on reversed-phase chromatography and tandem mass spectrometric (MS/MS) analysis of the peptides. As expected, Gln residues at different locations in the three-dimensional structure had different susceptibilities to deamidation. Gln deamidation was highly pH dependent with the highest level detected in the sample incubated at pH 9, and lowest level at pH 6 in the pH range from 5 to 9. The detection of significant levels of Gln deamidation suggested that it may play an important role in affecting heterogeneity and stability of recombinant monoclonal antibodies.

Assessment of antibody fragmentation by reversed-phase liquid chromatography and mass spectrometry.

Antibody fragmentation in the hinge region and other regions, and the impact of pH on the level and pattern of antibody fragmentation were investigated by reversed-phase (RP) liquid chromatography and mass spectrometry (LC-MS). Extensive fragmentation was observed in the hinge and in regions other than the hinge of a recombinant monoclonal antibody that was incubated in buffers of various pH at 40 degrees C for 10 weeks. Peptide bonds that were susceptible to hydrolysis were located mainly around the domain-domain interfaces close to or in the loop structures. The sites as well as the level of peptide bond hydrolysis were affected by the buffer pH. In agreement with previous findings when only the hinge region fragmentation was monitored, pH 6 was optimal for slowing down antibody fragmentation in regions other than the hinge. It also demonstrated that analysis by RPLC-MS provided a better assessment of the susceptible regions of recombinant monoclonal antibodies than size-exclusion chromatography (SEC) followed by fraction collection and mass spectrometry identification.

Effect of methionine oxidation of a recombinant monoclonal antibody on the binding affinity to protein A and protein G.

Oxidation of methionine (Met) residues is one of the most common protein degradation pathways. Two Met residues, Met256 and Met432, of a recombinant fully human monoclonal IgG1 antibody have been shown to be susceptible to oxidation. Met256 and Met432 are located in the antibody CH2-CH3 interface and in close proximity to protein A and protein G binding sites. The effect of oxidation of these susceptible Met residues on the binding to protein A and protein G was investigated in the current study. Incubation of the antibody with 5% tert-butyl hydroperoxide (tBHP) resulted in a nearly complete oxidation of Met256 and Met432, while incubation with 1% tBHP resulted in mixed populations of the antibody with different degrees of Met oxidation. Oxidation of Met256 and Met432 resulted in earlier elution of the antibody from protein A and protein G columns when eluted with a gradient of decreasing pH. Analysis by ELISA and surface plasmon resonance (SPR) revealed decreased binding affinity of the oxidized antibody to protein A and protein G. It is therefore concluded that oxidation of the Met256 and Met432 residues of the recombinant monoclonal antibody altered its interaction with protein A and protein G resulting in a decrease in binding affinity.

Fragmentation of a recombinant monoclonal antibody at various pH.

PURPOSE: To determine the relative importance of direct hydrolysis and beta-elimination, two common mechanisms of antibody hinge region fragmentation, and the impact of the conserved N-linked oligosaccharides in affecting antibody fragmentation under various pH. METHODS: A recombinant monoclonal antibody was incubated in buffers of various pH at 40 degrees C for 5 weeks. The level of fragmentation was measured using size-exclusion-chromatography (SEC). The specific sites of fragmentation were determined by analyzing SEC fractions using liquid chromatography mass spectrometry (LC-MS). RESULTS: Direct hydrolysis was accelerated by acidic and basic pH, while beta-elimination contributed to hinge region fragmentation at pH 7 and above. In addition, a shift of the major peptide bond hydrolysis sites in the hinge region towards the C-terminal direction with the decrease of sample pH from 9 to 5 was observed. At pH 4, the major cleavage site shifted outside the hinge region and was localized in the CH2 domain. Oligosaccharides did not affect hinge region fragmentation in the pH range of 5-9, however, at pH 4 oligosaccharides slowed down fragmentation in the CH2 domain. CONCLUSIONS: Antibody fragmentation level, sites and mechanisms were affected by pH. Oligosaccharides only affected the rate of fragmentation at pH 4.

Characterization of the glycosylation state of a recombinant monoclonal antibody using weak cation exchange chromatography and mass spectrometry.

Recombinant monoclonal antibody heterogeneity is inherent due to various enzymatic and non-enzymatic modifications. In this study, a recombinant humanized monoclonal IgG1 antibody with different states of glycosylation on the conserved asparagine residue in the CH(2) domain was analyzed by weak cation exchange chromatography. Two major peaks were observed and were further characterized by enzymatic digestion and mass spectrometry. It was found that this recombinant monoclonal antibody contained three glycosylation states of antibody with zero, one or two glycosylated heavy chains. The peak that eluted earlier on the cation exchange column contained antibodies with two glycosylated heavy chains containing fucosylated biantennary complex oligosaccharides with zero, one or two terminal galactose residues. The peak that eluted later from the column contained antibodies with either zero, one or two glycosylated heavy chains. The oligosaccharide on the antibodies eluted in the later peak was composed of only two GlcNAc residues. These results indicate that conformational changes in large proteins such as monoclonal antibodies, caused by different types of neutral oligosaccharides as well as the absence of oligosaccharides, can be differentiated by cation exchange column chromatography.

Structural effect of deglycosylation and methionine oxidation on a recombinant monoclonal antibody.

Methionine (Met) is one of the most susceptible amino acids to oxidation. Met256 (CH2-Met15.1) and Met432 (CH3-Met107) of a recombinant humanized monoclonal IgG1 antibody are located in the CH2 and CH3 domains, respectively. In three-dimensional structure, these two Met residues are close to the CH2-CH3 interface. In close proximity, oligosaccharides on the conserved asparagine (Asn) residues are enclosed in the CH2 domains. The relationship of Met oxidation with oligosaccharides and their effect on the structure of the antibody was investigated. Removal of oligosaccharides did not alter the oxidation rates of Met256 and Met432, however it caused significant structural changes as evidenced by the susceptibility of the deglycosylated antibody to trypsin and chymotrypsin. Oxidation of Met256 and Met432 did not cause significant conformational changes of the antibody with oligosaccharides, however oxidation of these Met residues accelerated degradation of the deglycosylated antibody. Analysis by mass spectrometry indicated that most of the protease cleavage sites were in the CH2 domains, which suggested that conformational changes induced by the removal of oligosaccharides and further by Met oxidation were local to the CH2 domains.

Mass spectrometry analysis of in vitro nitration of a recombinant human IgG1 monoclonal antibody.

Nitration of a recombinant human monoclonal antibody was carried out in vitro by incubating the antibody with the nitrating reagent tetranitromethane (TNM). The susceptible sites of nitration were identified using high-performance liquid chromatography/mass spectrometry (HPLC/MS). In general, tyrosine residues in the variable domains of the antibody are more susceptible to nitration, while tyrosine residues in the constant domains are relatively resistant to nitration. However, one tyrosine residue in the CH1 domain and one tyrosine residue in the CH2 domain are highly susceptible to nitration. Interestingly, the susceptible tyrosine residue in the CH2 domain is followed by the conserved asparagine residue that is glycosylated.

Heterogeneity of monoclonal antibodies.

Heterogeneity of monoclonal antibodies is common due to the various modifications introduced over the lifespan of the molecules from the point of synthesis to the point of complete clearance from the subjects. The vast number of modifications presents great challenge to the thorough characterization of the molecules. This article reviews the current knowledge of enzymatic and nonenzymatic modifications of monoclonal antibodies including the common ones such as incomplete disulfide bond formation, glycosylation, N-terminal pyroglutamine cyclization, C-terminal lysine processing, deamidation, isomerization, and oxidation, and less common ones such as modification of the N-terminal amino acids by maleuric acid and amidation of the C-terminal amino acid. In addition, noncovalent associations with other molecules, conformational diversity and aggregation of monoclonal antibodies are also discussed. Through a complete understanding of the heterogeneity of monoclonal antibodies, strategies can be employed to better identify the potential modifications and thoroughly characterize the molecules.

Characterization of lower molecular weight artifact bands of recombinant monoclonal IgG1 antibodies on non-reducing SDS-PAGE.

SDS-PAGE under non-reducing conditions is one of the most commonly used techniques for recombinant monoclonal antibody purity and stability indicating assay. On non-reducing SDS-PAGE, bands with a lower molecular weight than the intact antibody are routinely observed and is a common feature of IgG molecules. These fragments were analyzed by in-gel digestion followed by matrix-assisted-laser-desorption-ionization time-of-flight (MALDI-TOF) mass spectrometry, Western blot and by comparing the banding pattern of sample prepared in the presence of a reducing reagent. The fragments bands were identified as antibody lacking one light chain, two heavy chains, one light chain and one heavy chain, free heavy chain and free light chain. Sensitivity of fragmentation to sample buffer pH, incubation time, reducing reagent and alkylation reagents indicated that fragments were formed during sample preparation, but not present in the samples analyzed. Disulfide bond scrambling and beta-elimination are the two major mechanisms of the formation antibody fragments. Mass spectrometry analysis suggested that disulfide bond scrambling can be prevented by specifically modifying free sulhydryl using alkylation and thus reduced the amount of artifacts on non-reducing SDS-PAGE. Breakage of disulfide bonds by beta-elimination was evidenced by the detection of dehydroalanine using mass spectrometry.

Comparison of methionine oxidation in thermal stability and chemically stressed samples of a fully human monoclonal antibody.

Methionine (Met) oxidation is a major degradation pathway of protein therapeutics. Met oxidation of a fully human recombinant monoclonal antibody was investigated under both chemically stressed conditions using tert-butylhydroperoxide (tBHP) and thermal stability conditions where the sample was incubated in formulation buffer at 25 degrees C for 12 months. This antibody has one Met residue on each of the light chains and four Met residues on each of the heavy chains. In the thermal stability sample, only Met residues 256 and 432 in the Fc region were oxidized to form methionine sulfoxide, while Met residues in the Fab region were relatively stable. The susceptibility of Met residues 256 and 432 was further confirmed by incubating samples with tBHP, which has been shown to induce Met oxidation. Further analysis revealed that the susceptible Met residues of each heavy chain were randomly oxidized in samples incubated with tBHP, while in the thermal stability sample, the susceptible Met residues of one heavy chain were preferentially oxidized.

Characterization of the stability of a fully human monoclonal IgG after prolonged incubation at elevated temperature.

The susceptible degradation sites of therapeutic proteins are routinely assessed under accelerated conditions such as exposure to chemicals or incubation at elevated temperature or a combination of both. A fully human monoclonal IgG(1) antibody was characterized after incubation at 40 degrees C for 6 months by employing mass spectrometry and chromatography analyses. It was found that deamidation, fragmentation and N-terminal glutamate cyclization to form pyroglutamate are the major degradation pathways. Three major deamidation sites were identified and one site in a small tryptic peptide accounted for more than 80% of the total. Peptide cleavage was observed at several positions between different pairs of amino acids. Most of the cleavage sites were located in the hinge or other flexible regions of the IgG molecule.

A 13C NMR approach to categorizing potential limitations of alpha,beta-unsaturated carbonyl systems in drug-like molecules.

Compounds that contain an alpha,beta-unsaturated carbonyl moiety are often flagged as potential Michael acceptors. All alpha,beta-unsaturated carbonyl moieties are not equivalent, however, and we sought to better understand this system and its potential implications in drug-like molecules. Measurement of the (13)C NMR shift of the beta-carbon and correlation to in vitro results allowed compounds in our collection to be categorized as potential Michael acceptors, potential substrates for NADPH, or as photoisomerizable.