Abstract
Leachables are chemical compounds that migrate from manufacturing equipment, primary containers and closure systems, and packaging components into biopharmaceutical and pharmaceutical products. Acrylic acid (at concentration around 5 μg/mL) was detected as leachable in syringes from one of the potential vendors (X syringes). In order to evaluate the potential impact of acrylic acid on therapeutic proteins, an IgG 2 molecule was filled into a sterilized X syringe and then incubated at 45 °C for 45 days in a pH 5 acetate buffer. We discovered that acrylic acid can interact with proteins at three different sites: (1) the lysine side chain, (2) the N-terminus, and (3) the histidine side chain, by the Michael reaction. In this report, the direct interactions between acrylic acid leachable and a biopharmaceutical product were demonstrated and the reaction mechanism was proposed. Even thought a small amount (from 0.02% to 0.3%) of protein was found to be modified by acrylic acid, the modified protein can potentially be harmful due to the toxicity of acrylic acid. After being modified by acrylic acid, the properties of the therapeutic protein may change due to charge and hydrophobicity variations.
LAY ABSTRACT: Acrylic acid was detected to migrate from syringes (Vendor X) into a therapeutic protein solution (at a concentration around 5 μg/mL). In this study, we discovered that acrylic acid can modify proteins at three different sites: (1) the lysine side chain, 2) the N-terminus, and 3) the histidine side chain, by the Michael reaction. In this report, the direct interactions between acrylic acid leachable and a biopharmaceutical product were demonstrated and the reaction mechanism was proposed.
Introduction
Leachables are chemical compounds that migrate from manufacturing equipment, primary containers and closure systems, and packaging components into biopharmaceutical and pharmaceutical products during processing, storage, and shipping. Leachables can potentially cause toxicity of drug products and endanger patient safety (1). Compared to small molecule drugs, protein therapeutics may be more affected by leachables due to their large size and surface area. Leachables in protein therapeutics may pose more threats to patients because most of these drugs are administrated subcutaneously or intravenously. Although previous reports have discussed that leachables can cause degradation of therapeutic proteins (2–4) or interact with small molecule drugs (5), no direct interactions between leachates and proteins were identified. In this report, the direct interactions between acrylic acid leachable and a biopharmaceutical product are demonstrated and the reaction mechanism is proposed.
Acrylic acid is a component of the acrylic adhesive used to attach the needle to the barrel of pre-filled glass syringes. Even though acrylic acid was not detected in the syringes currently used by Amgen, it was identified as a leachable (∼5 μg/mL) in syringes from one of the potential vendors (X syringes). To investigate the potential interaction between the acrylic acid leachable and our protein drug products, a model IgG 2 antibody was filled into sterilized X syringes. After incubation, the antibody was digested and analyzed using an improved trypsin peptide mapping method (6).
Materials and Methods
Materials
The model IgG2 antibody was produced and purified at Amgen and kept at 4 °C in a pH 5.2 acetate buffer until use. tris(hydroxymethyl)aminomethane (TRIS), dithiothreitol (DTT), ammonium bicarbonate, and iodoacetic acid (IAA) were purchased from Sigma-Aldrich (St. Louis, MO). Water and acetonitrile (ACN) were obtained from VWR International (West Chester, PA). Formic acid (FA) and trifluoroacetic acid (TFA) was obtained from Pierce (Rockford, IL). Guanidine HCl (GdnHCl) was obtained from Mallinckrodt Baker (Phillipsburg, NJ). Lyophilized trypsins were obtained from Roche Applied Science (Indianapolis, IN).
Incubation
The model IgG 2 antibody at 15 mg/mL was filled into sterilized syringes X and then incubated at 45 °C for 45 days in a pH 5 acetate buffer.
Reduction, Alkylation, and Trypsin Digestion
A 15 mg/mL antibody was diluted to 1 mg/mL in 0.5 mL of pH 7.5 denaturation buffer (7.5 M guanidine HCl, 0.25 M TRIS). Reduction was accomplished with the addition of 3 μL of 0.5 M dithiothreitol followed by 30 min incubation at room temperature. Carboxymethylation was achieved with the addition of 7 μL of 0.5 M iodoacetic acid. The reaction was carried out in the dark for 15 min at room temperature. Excess iodoacetic acid was quenched with the addition of 4 μL of 0.5 M dithiothreitol. Reduced and alkylated samples were buffer exchanged into a pH 7.5 digestion buffer (0.1 M TRIS) using a NAP-5 column (GE Healthcare, Piscataway, NJ). Lyophilized trypsin was dissolved in water to a final concentration of 1 mg/mL. Digestion was started with the addition of the 1 mg/mL trypsin solution to the reduced, alkylated, and buffer-exchanged samples to achieve a 1:25 enzyme-to-substrate ratio. Digestion was carried out at 37 °C for 30 min. The final digest was quenched with the addition of 5 μL of 20% formic acid.
Reversed-Phase Liquid Chromatography (LC)
Reversed-phase separation of antibody digests was carried out on an Agilent 1200 series system (Agilent, Santa Clara, CA) equipped with a Polaris Ether 3 μm C18 2 × 250 mm column (Varian, Palo Alto, CA). Solvent A consisted of 0.1% aqueous TFA, and solvent B included 90% acetonitrile and 0.085% aqueous TFA. The temperature was maintained at 50 °C and the flow-rate was at 0.2 mL/min. A linear gradient from 0 to 50% B was run over 195 min.
Mass Spectrometry
A Thermo Finnigan (San Jose, CA) LTQ-Orbitrap mass spectrometer was used in-line with the high-perfomance liquid chromatography (HPLC) system to analyze the protein digests. A full MS scan followed by three data-dependant tandem mass spectrometry (MS/MS) scans were set up to acquire both the mass and the sequence information for the peptide maps. The full scan was collected with a resolution of 30,000 in the Orbitrap, and the MS/MS scans were acquired in the LTQ. The spray voltage was 5 kV and the capillary temperature was 300 °C. The instrument was tuned using the doubly charged ion of a synthetic peptide, Bradykinin. The MS/MS spectra were obtained using a normalized collision energy of 35%.
Results
Four tryptic peptides from the model IgG 2 antibody were observed to be covalently modified by acrylic acid (modification percentages varied from 0.02% to 0.3%). Acrylic acid was attached to four peptides via three different binding sites: (1) lysine side chains; (2) the N-terminus; and (3) histidine side chains. All the modifications were confirmed by acrylic acid–spiked experiments at 250 μg/mL concentration (see Appendix).
The first type of interaction between acrylic acid and the IgG 2 antibody was through the side chain of lysine. Figure 1(A) shows the selected ion chromatograms (SICs) of the unmodified peptide VDIKR (top) and its corresponding modified peptide (bottom). Compared to the native peptide, the modified peptide eluted later in the reversed-phase chromatogram because the additional acrylic acid increased the hydrophobicity of the peptide. The accurate mass measurements for both modified and unmodified peptides were within 3 ppm of the theoretical masses. The formulae of the doubly charged modified and the unmodified peptides were consistent with C30H57O10N9 and C27H53O8N9, respectively. The difference between these two formulae is C3H4O2, which is the formula for acrylic acid. This indicated that one acrylic acid is attached to the peptide. The SIC intensity of the modified peptide is about 0.3% of the native peptide. The linkage between the acrylic acid and the peptide was identified as a covalent bond because the acrylic acid unit was conserved after the collisionally induced fragmentation inside the mass spectrometer. By comparing the MS/MS spectra of the unmodified peptide (Figure 2A, top) and the modified peptide (Figure 2A, bottom), the binding site of acrylic acid was found to be on the lysine residue. The mass of y1 ion is the same for both modified and unmodified peptides. However, the masses of the y2, y3, and y4 ions for the modified peptide are 72 Da higher than the corresponding y ions of the unmodified peptide. This observation demonstrates that the acrylic acid is attached to the second residue from the C-terminal of the peptide, which is a lysine residue. The amino group on the lysine side chain was the most likely reaction site for the acrylic acid. Scheme 1A shows the proposed reaction mechanism between the lysine side chain and acrylic acid, which is a typical Michael addition reaction. Similar reactions were observed between the lipid product 4-hydroxy-2-nonenal (HNE) and proteins (7). In peptide VDIKR, the modified lysine residue was located in the middle of the tryptic peptide and was missed by analysis after trypsin digestion. This indicates that the acrylic acid modification on the lysine side chain prevents trypsin cleavage.
Selected ion chromatograph (SIC) for unmodified peptides (top) and the corresponding modified peptides (bottom). The acrylic acid modifications are through (A) lysine side chain, (B) N-terminus, and (C) histidine side chain. The amino acid sequences are labeled on the top of each chromatogram.
MS/MS spectra of unmodified peptides (top) and the corresponding modified peptides (bottom). The acrylic acid modifications are through (A) lysine side chain, (B) N-terminus, and (C) histidine side chain.
The second type of reaction between acrylic acid and the IgG 2 antibody was through the amino group at the N-terminus of the IgG 2 light chain. The N-terminal glutamine of the model IgG 2 antibody heavy chain was completely converted to pyroglutamate, which prevented the N-terminus amino group from interacting with acrylic acid. Hence, no modification was observed on the N-terminus peptide of the heavy chain. Figure 1B shows the SICs of the unmodified N-terminus light chain peptide DIVMTQTPLSSPVTLGQPASISCR (top) and its corresponding modified peptide (bottom). The modified peptide eluted later in the reversed-phase chromatogram and has an additional C3H4O2 unit in the formula (C111H189O40N29S2 vs C108H185O38N29S2) for the doubly charged ion compared to the unmodified peptide. These observations indicate an acrylic acid was attached to the modified peptide. Figure 2B shows the MS/MS data for both the modified (bottom) and unmodified (top) N-terminus peptides of the light chain. The y ions from y3 to y18 were singly charged and y192+ to y222+ were doubly charged. All the y ions up to y22 observed in the modified peptide were the same as their corresponding y ions from the unmodified peptide, which indicates that the modification did not occur within 22 residues from the C-terminus. As this peptide had only 24 residues, the modification should be on one of the first two residues, I (isoleucine) or D (aspartic acid). Both isoleucine and aspartic acid side chains are unlikely to interact with acrylic acid, so the modification of the N-terminus group by the acrylic acid is the most favorable reaction. As shown in Figure 2B, all the observed b ions from the modified peptide were 72 Da higher than that of the corresponding b ions of the unmodified peptide, which confirms that the modification is on the N-terminus. Scheme 1B is the proposed reaction between the N-terminus amino group and acrylic acid.
Proposed mechanisms for the interactions between acrylic acid and (A) lysine side chain, (B) N-terminus, and (C) histidine side chain.
The last type of modification was through histidine side chains. Two peptides were observed to be modified by acrylic acid through the histidine residues in our study. Figure 1C shows the SICs of the unmodified peptide VVSVLTVVHQDWLNGK (top) and its corresponding modified peptide (bottom). Interestingly, there are two isomers for the modified peptide (other modified peptides through histidine also had two isomers; data not shown). These two isomers had identical masses and MS/MS spectra, which indicate that the acrylic acid modifications for both isomers were at the same residue. The difference in formula between modified and unmodified peptides (C85H138O25N22 vs C82H134O23N22) was a C3H4O2 unit, which is the formula of acrylic acid. Figure 2C shows the MS/MS spectra for the modified peptide (bottom) and its corresponding unmodified peptide (top). The masses of y5, y6, and y7 ions from the modified peptide were the same as the masses of the corresponding y ions from the unmodified peptide, while the y8 to y13 ions from the modified peptide were 72 Da higher compared to those y ions from the unmodified peptide. These observations unambiguously identified the eighth residue from the C-terminus of the peptide, a histidine, as the site of the acrylic acid modification. The masses of the b9 to b15 ions for the modified peptide were 72 Da higher than the corresponding b ions from the unmodified peptide, which also supports the argument that the modification was on histidine. Scheme 1C is the proposed reaction between acrylic acid and the histidine side chain. The imidazole ring of histidine has two tautomers (Scheme 1C), with the tautomer I being more dominant (8). These two tautomers can react with acrylic acid to produce two isomers of the modified peptides. Although the two tautomers of histidine cannot be separated by the reversed-phase LC method, after being modified by acrylic acid the isomers can be baseline-separated using the same LC method. Because the histidine tautomer I is more dominant, its acrylic acid adduct is likely to be more abundant than tautomer II. Hence the larger peak eluting later in the reversed-phase chromatogram (Figure 1C, bottom) is expected to be tautomer I and the smaller peak eluting earlier is expected to be tautomer II.
Discussion
We discovered that acrylic acid can interact with therapeutic proteins at three different sites, lysine, the N-terminus, and histidine, by the Michael reaction. In our study, the concentration of the model protein was 15 mg/mL while the concentration of the leachable acrylic acid was relatively low (∼5 μg/mL). Hence just a small percentage of the protein (0.02–0.3%) was modified by the acrylic acid leachable due to the low leachable/protein ratio. It is reasonable to assume that therapeutic biologics with lower protein concentration may be affected to a greater extent because the leachable-to-protein ratio is more favorable to adducts formation. Increased acrylic acid concentration will also cause more protein modification. In the spiking experiment, 250 μg/mL acrylic acid was incubated with 15 mg/mL model protein. Besides four modified peptides presented above, another six new peptides were found to be modified by acrylic acid. The percentage of modified peptides was also found to be higher (up to 5%; see Appendix). It is worth mentioning that the amount of acrylic acid leachable will continuously increase with time because more and more acrylic acid will leach into protein solution from syringe in the process of storage. Hence, it is expected that more protein will be modified in long-term storage.
Another consideration is the effect of pH on this reaction. In our study, model protein was formulated in a pH 5 buffer. No experiments at high pH were conducted due to the instability of the model protein. However, it is reasonable to expect high pH will favor the formation of protein-leachable adducts because the interaction between acrylic acid and protein is a typical Michael Addition, a nucleophilic reaction. With increased pH, more free amino groups are available, which is more favorable for nucleophilic attack (9, 10).
After being modified by acrylic acid, the properties of the therapeutic protein may change significantly. This modification converts positively charged amino acid residues (lysine, histidine, and the N-terminus) into negatively charged groups, which can affect the isoelectric point and surface charge distribution of the protein and thus affect protein-protein interactions. The hydrophobicity of the protein is also expected to change due to the addition of acrylic acid. The changes in charge and hydrophobicity may lead to protein structure and functionality changes (11, 12). Obviously, the interactions between leachables and biopharmaceutical products should be avoided to ensure product quality.
Conflict of Interest Declaration
The authors declare that they have no competing interests.
Acknowledgments
We thank Jinkun Huang and Professor Sunny Zhou from Northeastern University for useful discussions, and Gary Rogers, Linda Narhi, and Joseph Phillips for suggestions and critical reading of the manuscript.
Appendix
Spiking Experiment Detail
The mixture of model IgG 2 antibody at 15 mg/mL and acrylic acid at 250 μg/mL was incubated at 37 °C for 15 days in a pH 5 acetate buffer in glass vials. After incubation, the antibody was digested and analyzed using an improved trypsin peptide mapping method as is described in experimental part in the main text.
Spiking Experiment Results
Ten peptides were observed to be modified by acrylic acid (beside four peptides observed in prefilled syringes, another six new peptides were modified). Five peptides were modified through side chain of lysine, one peptide through N-terminus, and four peptides through side chain of histidine. The modification percentage was varied from 0.2% to 5.0%.
Those four modified peptides observed in prefilled syringes were confirmed by the spiking experiments. The selected ion chromatograph (SIC) for unmodified peptides (top) and the corresponding modified peptides (bottom) via spiking experiment is shown in Figure 1. Figure 2 shows MS/MS spectra of unmodified peptides (top) and the corresponding modified peptides (bottom) via spiking experiment.
Selected ion chromatograph (SIC) for unmodified peptides (top) and the corresponding modified peptides (bottom) via spiking experiment. The acrylic acid modifications are through (A) lysine side chain, (B) N-terminus, and (C) histidine side chain. The amino acid sequences are labeled on the top of each chromatogram.
MS/MS spectra of unmodified peptides (top) and the corresponding modified peptides (bottom) via spiking experiment. The acrylic acid modifications are through (A) lysine side chain, (B) N-terminus, and (C) histidine side chain.
- ©PDA, Inc. 2012
Jump to section
Related Articles
Cited By...
- Holistic Extractables and Leachables Program: Evaluations of Prefilled Syringe Systems for Biotechnology Products
- A Comparison of Protein Stability in Prefillable Syringes Made of Glass and Plastic
- Characterization of Protein Aggregating Tungstates: Electrospray Mass Spectrometry Analysis of Extracts from Prefilled Syringes and from Tungsten Pins Used in the Manufacture of Syringes
- Creating a Holistic Extractables and Leachables (E&L) Program for Biotechnology Products
- Extractables Analysis of Single-Use Flexible Plastic Biocontainers