Abstract
Aspartate (Asp) isomerization is a common post-translational modification of recombinant therapeutic proteins that can occur during manufacturing, storage, or administration. Asp isomerization in the complementarity-determining regions of a monoclonal antibody may affect the target binding and thus a sufficiently robust quality control method for routine monitoring is desirable. In this work, we utilized a liquid chromatography–mass spectrometry (LC/MS)-based approach to identify the Asp isomerization in the complementarity-determining regions of a therapeutic monoclonal antibody. To quantitate the site-specific Asp isomerization of the monoclonal antibody, a UV detection–based quantitation assay utilizing the same LC platform was developed. The assay was qualified and implemented for routine monitoring of this product-specific modification. Compared with existing methods, this analytical paradigm is applicable to identify Asp isomerization (or other modifications) and subsequently develop a rapid, sufficiently robust quality control method for routine site-specific monitoring and quantitation to ensure product quality. This approach first identifies and locates a product-related impurity (a critical quality attribute) caused by isomerization, deamidation, oxidation, or other post-translational modifications, and then utilizes synthetic peptides and MS to assist the development of a LC-UV–based chromatographic method that separates and quantifies the product-related impurities by UV peaks. The established LC-UV method has acceptable peak specificity, precision, linearity, and accuracy; it can be validated and used in a good manufacturing practice environment for lot release and stability testing.
LAY ABSTRACT: Aspartate isomerization is a common post-translational modification of recombinant proteins during manufacture process and storage. Isomerization in the complementarity-determining regions (CDRs) of a monoclonal antibody A (mAb-A) has been detected and has been shown to have impact on the binding affinity to the antigen. In this work, we utilized a mass spectrometry–based peptide mapping approach to detect and quantitate the Asp isomerization in the CDRs of mAb-A. To routinely monitor the CDR isomerization of mAb-A, a focused peptide mapping method utilizing reversed phase chromatographic separation and UV detection has been developed and qualified. This approach is generally applicable to monitor isomerization and other post-translational modifications of proteins in a specific and high-throughput mode to ensure product quality.
- Monoclonal antibody (mAb)
- Complementarity-determining regions (CDRs)
- Aspartic acid isomerization
- Focused peptide mapping
- Method qualification
Introduction
The non-enzymatic isomerization of aspartic acid (Asp) residues in proteins is one of the major protein degradation pathways (1, 2) both in vitro (3, 4) and in vivo (5). In a mildly acidic buffer, Asp residues in a protein may isomerize spontaneously when the amide backbone nitrogen has a nucleophilic attack on the side chain carbonyl carbon to form an unstable, five-member succinimide ring. The resulting succinimide ring then undergoes rapid hydrolytic cleavage at the α or β carbon to form iso-Asp or Asp products, respectively (see Figure 1), with a molar ratio of 3:1 (6).
Asp isomerization has been studied extensively in both peptide (7) and protein models (4). Overall, the main factors affecting Asp isomerization are primary structure, higher-order structure (which determines the exposure to solvent and buffer components and conformational flexibility) (8), type of buffer used, pH, and temperature. Asp isomerization is prone to happen at the Asp residue in the motifs of Asp–Gly, Asp–Ser, Asp–Thr, and Asp–Asp (9⇓–11) at solvent-accessible regions. Recombinant proteins with these Asp isomerization “hot spots” are susceptible to Asp isomerization during manufacturing and storage of protein-based pharmaceuticals. Most of the formulation conditions chosen to minimize modifications such as deamidation, oxidation, and aggregation (e.g., a pH ranging from 4.5 to 7.2) cannot prevent the occurrence of Asp isomerization formation (4). Isomerization is considered one of the major post-translational modifications contributing to degradation of biopharmaceuticals (12).
Isomerization of Asp residues in a protein results in the formation of iso-Asp variants of the native protein molecule, which have an extra methylene group, originally from the side chain, inserted into the polypeptide backbone and a re-oriented and shortened acidic side chain. The Asp residues in complementarity-determining regions (CDR) of antibodies likely undergo isomerization because CDR regions are structurally flexible and solvent-accessible for target binding. There have been numerous reports about CDR Asp isomerization in monoclonal antibodies (mAbs). Cacia et al. (3) have reported that, upon Asp isomerization at the heavy chain complementarity-determining region 2 (CDR2), the relative binding affinity of a mAb binding to immunoglobulin E (IgE) was reduced by 87%. In another report, the binding of a mAb to epidermal growth factor receptor was abolished due to the avidity loss originating from the isomerization of a single Asp residue at the light chain CDR3 (13). Apart from causing protein conformational changes and loss of activity, Asp isomerization may also introduce or enhance immunogenicity as demonstrated in the T cell proliferation assays (14).
Given these reports, it is important to monitor Asp isomerization in the protein CDR region during the development of therapeutic protein products to ensure lot-to-lot consistency, clinical efficacy, and safety. The detection of iso-Asp variants is challenging, and the development and implementation of quantitative assays are often time-consuming and laborious. This is mostly due to the fact that Asp isomerization does not change the mass or net charge of a protein (9). The detection of iso-Asp variants in a recombinant protein is often performed by various chromatographic techniques in conjunction with N-terminal sequencing and/or mass spectrometry (MS). Cacia et al. (3) and Wakandar et al. (4) used hydrophobic interaction chromatography (HIC) to separate papain derived iso-Asp-containing protein fragment and identified the sites using N-terminal sequencing and liquid chromatography–mass spectrometry (LC-MS). In another report, the iso-Asp light chain variant could be separated by reversed-phase (RP) high-performance liquid chromatography (HPLC) and then further characterized with size exclusion chromatography (13). Cournoyer et al. (15) and O' Connor et al. (16) have demonstrated that isoaspartyl residues could be identified by electron capture dissociation and electron transfer dissociation MS, respectively. An 18O water–labeling approach in conjunction with chromatography techniques has also been used to detect Asp isomerization when there is accumulation of a succinimide intermediate (8⇓–10). Assays to quantitate Asp isomerization in protein include total isomerization measurement and site-specific measurement. The total Asp isomerization measurement uses a combination of enzyme treatment and HPLC-based quantitation. This approach is well known and often referred to as Isoquant® (Promega, Madison, WI). It utilizes the enzyme isoaspartyl methyl transferase to catalyze the transfer of a methyl group from S-adenosyl-L-methionine to iso-Asp at the α-carboxyl position, generating S-adenosyl homocysteine (SAH) in the process (17). The resulting SAH is then quantitated by reversed-phase (18) or ion exchange chromatography (19, 20). The limitation of Isoquant is that it is (1) unable to differentiate iso-Asp generated by Asn-deamidation and Asp isomerization, and (2) difficult to correlate the measured value to a specific isomerization site, as the measurement represents the total iso-Asp resulting from multiple deamidation and isomerization sites (this is often the case for large therapeutic proteins such as mAbs).
The second approach of quantitation measures Asp isomerization based on the separation of the variants at protein fragment/peptide level (i.e., the iso-Asp-containing large fragment, or peptide) (4, 10, 21, 22). In a recently published paper, Eakin et al. (22) compared results from two UV-based analytical methods (digested HIC and UV-focused peptide map) with the mass spectrometric result for Asp isomerization quantitation on one Asp site in the CDR region of a mAb. Even though the overall trend in iso-Asp content is similar in these three methods, there are consistent offset between the results from the two UV-based methods and the MS results—the digested HIC results are all lower than the MS results, while the focused peptide map results are all higher than the MS results. The exact cause for these discrepancies was unknown to the authors, but they suspected a lower recovery from the digested HIC method and co-elution of other species in the iso-Asp peak in the focused peptide map method.
In this study, we developed a streamlined, two-step analytical approach to identify and then to quantitate the Asp isomerization in a mAb. The first step was to profile the potential sites, and confirm isomerization sites using RP-LC-MS and synthetic peptides. The second step was to quantitate the site-specific isomerization by optimizing the RP–ultra performance liquid chromatography (UPLC) method to ensure baseline separation of peaks related to isomerization sites of interest. The applicability of this method is demonstrated by identifying and then quantitating multiple isomerization sites of a therapeutic mAb. We found good agreement of isomerization percentages between data from UV and from MS, owing to the high purity of the UV peaks used in quantitation. The RP-UPLC method was subsequently qualified and found to be accurate and specific, as well as rapid. It's suitable for both characterization and routine testing. Without the need of using MS for quantitation, this LC-UV based quantitation method can be easily implemented in a regulated environment (23).
Materials and Methods
Materials
MAb-A, which is an immunoglobulin G1, was produced by MedImmune (Gaithersburg, MD). Trypsin was from Calbiochem (San Diego, CA). Dithiothreitol (DTT, no-weight format) was from Pierce (Rockford, IL). Urea (OmniPur®), water (OmniSolve, HPLC and spectrophotometry grade) and acetonitrile (OmniSolve, HPLC and spectrophotometry grade) were from EMD (Billerica, MA). Iodoacetamide (IAA; OneQuantTM) was from G Biosciences. Trifluoroacetic acid (TFA) in flame-sealed, 1 mL ampules was from Sigma (St. Louis, MO).
Peptides H6 and L7 are the 6th and 7th mAb-A tryptic fragments (from the N-terminus) of the heavy chain that constitutes the CDR2, and light chain that constitutes the CDR3, respectively. Iso-H6 and iso-L7 are the iso-Asp substituted counterparts for H6 and L7 (see Table I). Synthetic H6, iso-H6, L7, and iso-L7 were made by the core facility of the Physiology Department, Tufts University (Medford, MA).
Stability Studies
MAb-A at a concentration of 20 mg/mL in 10 mM histidine, 2.35% lysine hydrochloride buffer, and 0.02 polysorbate 80 at pH 6.0 was placed at 5 °C, 25 °C, and 40 °C. Samples were tested at different time points.
Potency Assay
Biological activity of mAb-A was measured relative to a reference material using a cell-based reporter gene assay. This assay measures the ability of mAb-A to neutralize the human interferon-α2b receptor (IFNαR) through the prevention of the binding of interferon-α2b (IFN) ligand to the receptor. The bioactivity was measured using a HEK-293H cell line transfected with a luciferase reporter gene driven by an interferon stimulated response element. When IFNα binds to its receptor on the cell surface, stimulation of the signaling pathway activates luciferase expression that is inversely proportional to the activity of mAb-A. The relative potency is determined by dividing the half maximal inhibitory concentration (IC50) of the reference standard by the IC50 of each test sample and expressed as a percentage of the reference material.
Tryptic Digestion
Samples were diluted to 5 mg/mL, mixed with denaturing buffer 8 M guanidine, 130 mM Tris, 1 mM ethylenediaminetetraacetic acid (EDTA), pH 7.6 and 500 mM dithiothreitol (DTT), vortexed, and incubated at 37° C for 30 min. Iodoacetamide was added for alkylation, and the mixture was incubated in the dark at room temperature for 30 min. The reduced and alkylated samples were transferred to a dialysis cassette and dialyzed in 6 M urea and 100 mM Tris buffer (pH 7.6) for 2 h in the dark. Dialyzed samples were mixed with 100 mM Tris buffer (pH 7.6) and trypsin, and then incubated at 37 °C for 4 h. The digestion was quenched with TFA.
Preparing Synthetic Peptides for Assay Development
Because both L7 and iso-L7 contain a cysteine residue (see Table I), to match the L7 and iso-L7 derived from the tryptic digestion of mAb-A, synthetic peptides L7 and iso-L7 were alkylated prior to use. The peptides were first dissolved in 100 mM Tris (pH8.0) to 1 mg/mL. And then, 9 μL of 500 mM iodacetamide was added to 485 μL L7 or iso-L7 and the mixture was kept at room temperature in dark for 1 h. In the end, 7 μL of 500 mM DTT was added to quench the alkylation reaction. A mixture of all four synthetic peptides (H6, iso-H6, alkylated L7, and alkylated iso-L7) was diluted in water to a final concentration of 3.33 μM for each peptide.
LC-MS Instruments, Columns, and Methods
LC separation of peptides was evaluated on an ACQUITY UPLC® system (Waters, Milford, MA) or an Infinity UHPLC® system (Agilent, Santa Clara, CA). Peak identification by MS was carried out in a positive mode on Q-Tof Premier® (Waters) or Q-Tof Synapt® G1 (Waters) with MassLynx 4.1 software (Waters). Ethylene bridged hybrid (BEH)-300 C18 1.7 μm 2.1 × 150 mm column (24) (Waters part #186003687), charged surface hybrid (CSH) C18 1.7 μm 2.1 × 150 mm column (Waters part #186005298), and high strength silica (HSS) T3 1.8 μm 2.1 × 150 mm column (Waters part# 186003540) were evaluated with ultra-performance liquid chromatography (UPLC), and Zorbax rapid resolution high definition (RRHD) 300SB C18, 1.8 μm 2.1 × 100 mm column (Agilent part # 858750-902), as well as the custom-made Zorbax RRHD 300SB C18, 1.8 μm 2.1 × 150 mm column, were evaluated with ultra-high performance liquid chromatography (UHPLC).
For LC-MS based peptide mapping assay, the LC separation of peptides was performed on an ACQUITY UPLC® using a BEH-300 C-18 2.1 × 150 mm on an ACQUITY UPLC® at 55 °C with a flow rate of 200 μL/min, using 0.02% TFA in H2O as mobile phase A and 0.02% TFA in acetonitrile as mobile phase B. The tryptic peptides were eluted using mobile phase B (0–5% over 3 min, 5–20% over 24 min, 20–21% over 5.3 min, 21–24% over 4.2 min, 24–35% over 35 min, 35–90% over 4 min, 90–95% over 1 min, 95–5% over 1.5 min, 5–95% over 2 min, and 95–0% over 1 min).
The optimized LC separation was performed using a BEH300 C18 2.1 × 150 mm or Zorbax RRHD 300SB C18, 1.8 μm 2.1 × 150 mm column on an ACQUITY UPLC® or a UHPLC® at 55 °C with a flow rate of 200 μL/min, using 0.02% TFA in H2O as mobile phase A and 0.02% TFA in acetonitrile as mobile phase B. Peptides were eluted using mobile phase B (15–21% over 3.5 min, 21–23% over 6 min, 23% for 11.5 min, 23–90% over 1 min, 90% over 2 min, 90–15% over 0.1 min, and 15% over 6.9 min) at a flow rate of 0.2 mL/min. Eluted peptides were monitored by UV absorbance at 220 nm.
Results and Discussion
Detecting Asp-isomerized Variants
A mAb (mAb-A) was incubated in formulation buffer at 2–8 °C (storage temperature—unstressed condition) and 40 °C (stressed condition). Tryptic peptide mapping using RP separation followed by online mass spectrometric detection was used to monitor the levels of chemical modifications including Asp isomerization. Figure 2 is the extracted ion chromatograms of H6 and L7 from stressed and unstressed mAb-A. When searching the mass of H6 and L7, there were two pairs of peaks detected, with one pair having the mass of H6 and the other pair having the mass of L7. Tandem mass spectrometry (MS/MS) indicated that each pair of peaks has “identical” amino acid sequences. Moreover, the earlier eluting minor peaks in each pair grow in intensity under stressed conditions. The logical explanation for two peaks having the same mass, the same sequence, but different elution times is that these two peaks are isomers of the same peptide. Indeed, both H6 and L7 contain an Asp-Ser sequence motif that is susceptible to Asp isomerization formation (see Table I).
Because the MS/MS spectrum cannot differentiate an Asp from an iso-Asp, synthetic peptides of the native H6 and L7 as well as their isomers iso-H6 and iso-L7 were generated for peak confirmation. The experimental data showed synthetic iso-Asp H6 and iso-Asp L7 peptides have the identical retention time as each of the earlier eluting peaks in the two pairs of peptides in the tryptic digest of mAb A, confirming the earlier eluting peaks in each of the two pairs of peptides are iso-Asp variants of the H6 and L7 in the stability samples. Upon isomerization, the iso-Asp peptides become more hydrophilic due to the shortened acidic side chain (by a methylene group). Thus, iso-Asp peptides elute earlier in RP chromatography, which is consistent with previous reports that reduced hydrophobicity was observed on the iso-Asp variants (10, 13).
The succinimide intermediate (see Figure 1) can accumulate in heat-stressed mAb-A and has been identified in its ion exchange chromatography (IEC) basic peak 1 and 3 (data not shown). However, it was not detected in the tryptic peptide mapping analysis. This was understandable because tryptic peptide mapping was performed in slightly basic pH (7.6∼8.0) solution for optimal digestion. At this pH, the succinimide intermediate hydrolyzes rapidly, that is, mostly forming iso-Asp, partly reverting to Asp with a molar ratio of 3:1 (6), making it detected mostly as iso-Asp (8, 14).
Correlating CDR Asp Isomerization with Bioactivity
In this sudy for mAb-A, peptide mapping together with other physicochemical testing, such as size exclusion chromatography and gel electrophoresis, were used to monitor the potential degradants. The physicochemical methods did not indicate high levels of degradation such as aggregation or fragmentation during the stability study (See Supplementary Figure 1 for aggregation and fragmentation versus potency). However, an increase of Asp isomerization in the CDR region was observed and appeared to correlate with the potency loss measured by bioassay. Figure 3 is the plot of CDR Asp isomerization versus potency in the 25 °C and 40 °C stressed mAb-A. Under both conditions, cell-based bioactivity decreased when Asp isomerization variants increased. Besides CDR isomerization, Fc modifications including Fc deamidation, oxidation, and fragmentation were also observed to a lesser extent. Due to the fact that the effector function is not involved in the mechanism of action of mAb-A, these changes observed in the Fc region did not contribute to the loss of activity.
Antagonistic activity of mAb-A was characterized to the molecular basis by our colleagues Peng et al. (25). They found that mAb-A binds to a novel, function-blocking epitope on SD3 of interferon-a/β receptor 1 (IFNAR1) through two interfaces: a larger interface in which heavy chain (HC) CDR2 contributes most to the binding, and a small interface in which LC CDR3 contributes predominantly. By binding to the epitobe of IFNAR1 through these two interfaces, the LC CDR1 and framework 3 region of mAb-A partially occupy the space and inhibit IFN ligand binding to IFNAR1 receptor through steric hindrince. This indicates that a single arm of the mAb-A Fab would inhibit the IFNAR signaling pathway. However, with both arms of Fab, mAb-A will have higher apparent binding affinity due to avidity, which often results higher efficacy. Also notably, for both the larger interface in which HC CDR2 contributes most to the binding and the small interface in which LC CDR3 contributes predominantly, the contribution to binding by hydrophobic amino acids is small, with most of the binding conferred by charged residues (25). This aligns perfectly with our observation that isomerization of the charge residues Asp 55 in HC CDR2 and Asp 93 in LC CDR3 reduced bioactivity; therefore it is important to monitor the isomerization of these two sites.
Developing RP-UPLC Focused Peptide Map Method
As Asp isomerization formation of mAb-A occurs during the manufacture and storage process and affects bioactivity, it is important to develop and implement a UV detection–based focused peptide mapping assay to monitor this quality attribute for its ease of transferability to the QC lab.
Optimization of the Chromatography
Chromatographic separation of two isomeric peptides and their corresponding precursors from the rest of the mAb-A–derived tryptic peptides requires high-performance LC instruments in conjunction with high-resolution columns. A UPLC or UHPLC system with sub-2 μm particle-packed columns was selected to separate the two pairs of closely related peptides from a complex peptide mixture. The chromatography conditions of the previously described peptide mapping assay were used as the starting point for the assay development and optimization. A number of critical assay parameters and conditions, which include the LC instrument platforms, columns, temperature, flow rate, and TFA concentration, were evaluated as detailed in Table II. The selection of conditions and parameters of the assay is based on the specificity (resolution), overall run time, assay performance, and robustness. Certain experimental parameters (such as column temperature, column length, flow rate, and ion pairing reagent), which are important for the performance of chromatography separation, were the focus of the method optimization. Experimental conditions were selected with the goal of achieving the best purity of the two pairs of isomer peptides and the maximum resolution of these four peptides from their neighboring peptides. For example, we found column temperature affected the peak elution order and purity of the peptides as shown in Figure 4; compared to 60 °C and 65 °C, better resolution was achieved at 55 °C. Peak resolution affected by column length is demonstrated in Supplementary Figure 2, in which the 150 mm column showed a baseline separation compared to a 100 mm column of the same type.
The optimal separation of iso-H6, H6, iso-L7, and L7 from the rest of mAb-A tryptic peptides was achieved on both the UPLC and UHPLC system using BEH300 C18 1.7 μm 2.1 × 150 mm column or Zorbax RRHD 300SB C18, 1.8 μm 2.1 × 150 mm column (See Figure 5) at 55 °C with 0.02% TFA in water and acetonitrile using the gradient detailed in the Materials and Methods section. Both these two columns and systems showed good specificity, precision, and linearity; formal qualification work was performed on the UPLC system using a BEH column.
Confirmation of Peak Purity Using Mass Spectrometry
Throughout assay development, the identity and purity of the putative isomer peaks were confirmed with the mass spectrometric (MS) and tandem mass spectrometric (MS/MS) information. The MS spectra of iso-H6, H6, iso-L7, and L7 show that the peptides of interest are separated from each other and from the rest of the mAb-A tryptic peptides, and the amount of each peptide of interest can be quantitated by LC-UV peaks. The MS spectrum of each monitored peptide was “pure”, containing the expected peptide ion and in-source fragments of the same peptide (See Supplementary Figure 3). Synthetic peptides containing Asp or iso-Asp were used to identify/differentiate Asp and iso-Asp in mAb-A.
To verify that quantitation from the UV-based focused peptide map method is in good agreement with that from MS, isomerization percentages from UV and MS data were compared using following samples: mAb-A reference standard; mAb-A incubated in human serum at 37 °C for 3 weeks, 4 weeks, and 5 weeks (purified by protein A after incubation); and mAb-A stability sample at 40 °C for 2 months (See Figure 6). Although there was a large span of iso-Asp levels among these samples (from sub 2% to ∼34%), very narrow differences in isomerization percentages (0.1 to 1.8%) were observed between UV data and MS data on both isomerization sites. In addition, there was no constant offset between the UV data and MS data. These results are tighter than the results reported by Eakin et al. (22), in which UV data was constantly higher than MS data, and the highest difference was 4 to 5% among samples with isomerization levels ranging from sub 2 to ∼36%. The better agreement between our UV data and MS data should be attributed to the high purity of the UV peaks of interest from both isomerization sites.
Qualifying LC-based Focused Peptide Map Assay
Once the chromatographic conditions had been optimized and peak purity and identity confirmed by MS and MS/MS, the UV method was qualified to support its use as a stability-indicating assay in a good manufacturing practice (GMP) environment. Even though there is no clear definition or guideline of assay qualification, we qualified the method by evaluating several key characteristics of the assay as per the validation design described in the Guidelines of the International Conference on Harmonization (ICH, Q2R1). The ability of this method to quantify isomeric peptides by UV absorbance was evaluated by determining the assay's specificity, linearity, accuracy, precision, and limit of quantitation (LOQ) and limit of detection (LOD). Qualification results are summarized in Table III.
Specificity
The specificity was determined by analyzing a formulation blank, synthetic peptides mixture, and a tryptically digested 2 month stressed sample by LC-MS. As shown in Supplementary Figure 3, the MS spectrum of each monitored peptide was “pure”, containing the expected peptide and some in-source fragments of this peptide.
Precision
Precision (repeatability and intermediate precision) was evaluated using six replicates of the tryptically digested stressed sample by different days, analysts, and instruments. The percent coefficient of variation (% CV) of isomerization from the six replicates is less than 5% on either isomerization site.
Linearity and Range
For linearity, the matrix-matched calibration standards were obtained by spiking synthetic iso-Asp peptides (iso-H6 and iso-L7) into the tryptic digest of mAb-A reference material at 0.208, 0.417, 0.833, 1.67, and 3.33 μM. The assessment was performed in duplicate. Samples with isomer concentration in this range contain approximately 5 to 90% iso-H6 and approximately 3 to 56% iso-L7, presumably covering the Asp isomerization percentages ranging from undetectable activity loss to substantial or complete activity loss. Calibration curves were generated based on the isomer peak area relative to the spike concentration (see Figure 7). The R2 of the calibration curve is 0.9999 and 0.9982 for iso-H6 and iso-L7, respectively, demonstrating a good fit of the linear regressions.
Accuracy
For accuracy, synthetic iso-Asp peptides (iso-H6 and iso-L7) at three concentration levels (0.2, 1, and 2 μM), respectively representing low, middle, and high values on the linear curve, were spiked into the tryptically digested mAb-A in triplicate. The concentration of the spiked isomers was experimentally determined by interpolation of the peak area on the calibration curve after correcting for the analytes content already present in mAb-A (peak area for interpolation = isomer peak area in spiked sample – isomer peak area in un-spiked sample). Accuracy was calculated by comparing the known spiking concentration with the experimentally determined concentration. At all three spiking levels, the obtained recoveries were above 95% with the % CV below 10% on both isomerization sites, demonstrating high accuracy of the UV method.
Limit of Quantitation (LOQ) and Limit of Detection (LOD)
The LOQ and the LOD were evaluated from the calibration function. The LOQ was calculated as 10(SD/S) or 10(σ/S) and the LOD as 3.3(SD/S) or 3.3(σ/S), where SD or σ is the standard deviation of the y-intercepts and S is the slope of the calibration function (see regression data analysis in Supplementary Table I and Supplementary Table II). Our results demonstrate that the focused peptide mapping method can quantitate isomers at concentrations as low as 0.07 μM for iso-H6 and 0.26 μM for iso-L7. At these concentrations, isomerization percentage is approximately 2.6% for iso-6 and 4.0% for iso-L7, which is about the same isomer levels in mAb-A unstressed material (i.e., t = 0 in a stability study), and far below the levels that can incur demonstrable bioactivity loss.
System Suitability
System suitability was established by three standards: a mixture of four synthetic peptides, a tryptically digested mAb-A reference material, and a tryptically digested mAb-A stressed sample. The criteria include the expected UV profiles and a defined range of iso-H6 and iso-L7 percentages in each tryptic digested mAb-A standard.
Conclusions
In this study we developed a high-resolution, UV-based focused peptide map method to quantify Asp isomerization on two sites in CDR regions of a mAb. Our method is similar in principle to the method reported by Eakin et al. (22), but more challenging due to two CDR isomerization sites instead of one needing to be quantified. Nevertheless, a better agreement between UV data and MS data was achieved by our method compared to the method developed by Eakin et al. (22), thanks to the high purity of the UV peaks of interest in our method. This method was then successfully qualified and implemented as a stability-indicating assay easily transferable to a GMP environment. This assay has been proven to be specific and accurate to quantitate the Asp isomerization modifications and has been incorporated into the QC strategy for a therapeutic protein. Compared with other published methods for the identification and quantitation of previously unknown Asp isomerization, the methodology presented in this report is very streamlined and demonstrated good specificity, accuracy, low limit of quantitation, and good dynamic range on quantifying more than one isomerization site in one chromatographic run. Moreover, this LC-UV–based method can be validated and therefore used in the regulated environment.
Conflict of Interest Declaration
The authors declare that they have no competing interests.
SUPPORTING INFORMATION AVAILABLE
Acknowledgements
The authors thank Christina Ufholz, Liang Zhu, Hung-yu Lin, Jose Casas-Finet, David Spencer, Ken Miller, and Li Peng from MedImmune for their support of this work.
- © PDA, Inc. 2016
Reference
PDA members receive access to all articles published in the current year and previous volume year. Institutional subscribers received access to all content. Log in below to receive access to this article if you are either of these.
If you are neither or you are a PDA member trying to access an article outside of your membership license, then you must purchase access to this article (below). If you do not have a username or password for JPST, you will be required to create an account prior to purchasing.
Full issue PDFs are for PDA members only.
Note to pda.org users
The PDA and PDA bookstore websites (www.pda.org and www.pda.org/bookstore) are separate websites from the PDA JPST website. When you first join PDA, your initial UserID and Password are sent to HighWirePress to create your PDA JPST account. Subsequent UserrID and Password changes required at the PDA websites will not pass on to PDA JPST and vice versa. If you forget your PDA JPST UserID and/or Password, you can request help to retrieve UserID and reset Password below.