Enhanced Detection of Low-Abundance Host Cell Protein Impurities in

Aug 12, 2015 - The enormous dynamic range of proteinaceous species present in protein biotherapeutics poses a significant challenge for current mass s...
0 downloads 11 Views 1MB Size
Subscriber access provided by Georgetown University | Lauinger and Blommer Libraries

Article

Enhanced Detection of Low-Abundance Host Cell Protein (HCP) Impurities in High-Purity Monoclonal Antibodies Down to 1 ppm using Ion Mobility Mass Spectrometry coupled with Multidimensional Liquid Chromatography Catalin E. Doneanu, Malcolm Anderson, Brad J. Williams, Matthew Allen Lauber, Asish B Chakraborty, and Weibin Chen Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.5b02103 • Publication Date (Web): 12 Aug 2015 Downloaded from http://pubs.acs.org on September 3, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Enhanced Detection of Low-Abundance Host Cell Protein (HCP) Impurities in High-Purity Monoclonal Antibodies Down to 1 ppm using Ion Mobility Mass Spectrometry coupled with Multidimensional Liquid Chromatography

Catalin E. Doneanu1*, Malcolm Anderson2, Brad J. Williams3, Matthew A. Lauber1, Asish Chakraborty1 and Weibin Chen1*

1

Waters Corporation, 34 Maple Street, Milford, MA 01757, USA

2

Waters Corporation, Stamford Avenue, Altrincham Road, Wilmslow, SK9 4AX, UK

3

Waters Corporation, 100 Cummings Center, Suite 407N, Beverly, MA 01915, USA

*Corresponding authors: Phone: 508-482-3040 or 508-482-2090. Fax: 508-482-3085. E-mail: [email protected] or [email protected]

1 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract

The enormous dynamic range of proteinaceous species present in protein biotherapeutics poses a significant challenge for current mass spectrometry (MS)-based methods to detect lowabundance HCP impurities. Previously, an HCP assay based on two dimensional chromatographic separation (high pH/low pH) coupled to high-resolution quadrupole time-offlight (QTOF) mass spectrometry and developed in the author's laboratory, has been shown to achieve a detection limit of about 50 ppm (parts per milion) for the identification and quantification of HCPs present in monoclonal antibodies following Protein A purification [1]. To improve the HCP detection limit we have explored the utility of several new analytical techniques for HCP analysis and thereby developed an improved liquid chromatography-mass spectrometry (LC-MS) methodology for enhanced detection of HCPs. The new method includes (1) the use of a new charge-surface-modified (CSH) C18 stationary phase to mitigate the challenges of column saturation, peak tailing and distortion that are commonly observed in the HCP analysis; (2) the incorporation of travelling-wave ion mobility (TWIM) separation of coeluting peptide precursors and (3) the improvement of fragmentation efficiency of lowabundance HCP peptides by correlating the collision energy used for precursor fragmentation with their mobility drift time. As a result of these improvements, the detection limit of the new methodology was greatly improved, and HCPs present at a concentration as low as 1 ppm (1ng HCP/mg mAb) were successfully identified and quantified.

The newly developed method was applied to analyze two high-purity mAbs (NIST mAb and Infliximab) expressed in a murine cell line. For both samples, low-abundance HCPs (down to 1

2 ACS Paragon Plus Environment

Page 2 of 24

Page 3 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

ppm) were confidently identified, and the identities of the HCPs were further confirmed by targeted MS/MS experiments. In addition, the performance of the assay was evaluated by an inter-laboratory study in which three independent laboratories performed the same HCP assay on the NIST mAb sample. The reproducibility of this assay is also discussed.

3 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Introduction

In the field of small-molecule trace analysis, it is possible to measure very low levels of impurities down to ppb, ppt or even ppq levels [2-3]. This achievement is made possible by the fact that the sample matrix (e.g.

soil, food, water, etc) and the analyte of interest (e.g.

environmental pollutants) have drastically different chemical structures/properties, which allows for selective enrichment of the analyte. For a host cell protein (HCP) analysis, the exploration of HCP enrichment in order to reach lower levels of detection is not particularly desired nor easy to implement. Host cell protein impurities are most often present at low levels (1-100 ppm) in protein biopharmaceuticals and the matrix and the analyte are both proteinaceous and converted to peptides if the analysis is done via a "bottom-up" LC/MS approach. Without knowing the identity and properties of an HCP, there is no general sample preparation procedure that can enrich the analyte (HCP or HCP peptide) while removing the matrix background corresponding to the therapeutic. In other words, analytical techniques required for this application encounter the extreme challenge of dealing with about 1 million times more matrix molecules than the analyte. For this reason, a successful HCP assay needs to be able to deal with this very high sample complexity.

Because HCPs can sometimes elicit an immunogenic response, regulatory guidelines mandate that they be identified and quantified in order to protect patient safety [4]. LC/MS-based methods are becoming a common approach, where residual HCPs can be detected, identified, and quantified directly [5-16]. The utility of mass spectrometry-based HCP assays is recognized either in providing an orthogonal method to more traditional approaches (ELISAs) [17-19] or to

4 ACS Paragon Plus Environment

Page 4 of 24

Page 5 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

address some of their limitations [5,6]. Until recently, two-dimensional chromatographic approaches coupled to data-independent MS/MS acquisition have been exclusively employed to address the wide dynamic range and the complexity of HCP samples. These assays were able to achieve detection limits in the concentration range of 10-50 ppm within samples of moderate (low molecular weight biopharmaceuticals) or high complexity (therapeutic mAbs) [1,5-11]. Over the last two years, several reports using single dimension chromatographic separation in combination with data-dependent MS/MS acquisition have been presented [11-16]. With one exception [16], these assays rely on proteomics-scale nanoflow separations (performed on 75 µm ID long analytical columns using 3 h gradients) and are reported to achieve similar detection levels for HCP analyis (10-50 ppm). However, a single dimension chromatographic separation is particularly challenged in resolving the sample components across 5-6 orders of magnitude. Low-abundance HCP peptides frequently co-elute with high-abundance biopharmaceutical peptides. As a result, HCP peptides face ion suppression effects. Moreover, the reproducibility of HCP detection is deteriorated by the use of nanoflow columns under the heavily overloaded conditions (1-10 µg of mAb digest per injection) needed to achieve the reported HCP detection limits (10-50 ppm). For the development of a successful LC-MS assay for HCP analysis, one capable of achieving even lower limits of detection, multiple dimensions of separation seems to be preferred as a single chromatographic dimension struggles to offer the necessary dynamic range and peak capacity needed for such an assay.

In our previous work, we reported an HCP assay, using a system consisting of a 2D-high pH/low pH reversed phase (RP/RP) chromatographic separation coupled to high-resolution (~ 10,000) QTOF mass spectrometry. This instrumentation was consistently able to achieve ~50 ppm

5 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

detection limits for HCPs present in mAb samples (post-protein A purification) [1]. As a continuous effort to improve the detection limit of the LC-MS assay for HCP analysis, we have now exploited several key analytical metrics (e.g. sample loadability, system peak capacity, and optimization of fragmentation energy) to enhance the performance of this approach. In this work, we describe how we have incorporated several new technological enhancements into the HCP assay to greatly improve its detection limit. Two high-purity mAb samples, a NIST-sourced mAb and Infliximab (a licensed biopharmaceutical mAb) were analyzed by the newly developed method. A significant number of HCPs with concentrations ranging from 1-100 ppm were identified from the NIST mAb sample, and for Infliximab, two HCPs present in the 10-25 ppm range (epidermal growth factor-like protein 8 and WD repeat-containing protein 37) were also identified. In addition, an inter-laboratory study was performed at three sites to evaluate the reproducibility of the method for HCP analysis using the NIST mAb. Among all the HCPs identified by the three laboratories, more than 70% of HCPs were identified by two sites. Out of 14 HCPs identified by all three laboratories for the NIST sample, eleven HCPs in the concentration range of 1-30 ppm, were subsequently validated by acquiring "confirmatory" MS/MS fragmentation spectra. All of the results suggest that the 2D LC/IMS-MS method is a viable approach to detecting low abundance HCPs in high purity protein therapeutics.

Experimental

A diagram illustrating the configuration of the 2D-LC RP/RP system used for the HCP assay is shown in Figure 1. Full details regarding the experimental conditions used for sample preparation and for system operation are presented in the Supplementary Materials section).

6 ACS Paragon Plus Environment

Page 6 of 24

Page 7 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Results and Discussion

An essential and yet extremely challenging aspect of a successful HCP assay is its ability to detect relevant HCPs in a highly purified protein therapeutic. An HCP assay previously developed in our laboratory using comprehensive two-dimensional (2D) chromatography separations (high pH/low pH RP/RP) coupled to high-resolution (~ 10,000) QTOF mass spectrometry was able to identify and measure HCPs down to ~50 ppm [1]. The detected HCPs were from monoclonal antibody (mAb) samples that were harvested from CHO (chinese hamster ovary) cell culture media and underwent a single Protein A affinity purification. In this report, we explored the analytical basis supporting the performance of LC-MS methodologies for detecting low abundance HCPs in highly-purified biophamaceuticals. Three major modifications were made to the original assay aimed at improving the HCP detection limit: 1) we increased the amount of sample that can be loaded onto the second dimension (low pH) separation by using a column packed with a charge-surface-modified C18 material (CSH); 2) we employed a travelling-wave ion mobility separation of peptide precursors and 3) we used a quasi "fixed" collision energy (CE) for fragmentation of peptide ions, correlated with the mobility drift time of their precursors. Each of these new developments addresses a very specific limitation of the original HCP assay. The capability to load five times more sample (500 µg vs 100 µg mAb digest) is directly related to the use of charge-surface-modified C18 and being able to minimize the peak distortion that is characteristic of "conventional" C18 stationary phases when they are subjected to high-mass loads under low pH conditions. It is well documented that ionic analytes (e.g. peptides) exhibit strong interactions (either solute-solute, solute-sorbent, or a combination of both) in low pH/low ionic strength LC mobile phases. These type of interactions are very

7 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

likely responsible for the poor loadability of peptides, which is especially exacerbated at high concentrations [20-23]. In fact, in bottom- up HCP assays it is common to observe the biopharmaceutical derived peptides eluted as very broad, tailing, split, and/or "double" peaks (which can be an indication of partial loss of peptide retention). To accommodate the loadability requirement, a relatively wide-bore, (1 mm ID column) has been utilized as the first dimension of the 2D RP/RP setup shown in Figure 1. Such a column can be loaded with high sample amounts (e.g. 200-1,000 µg) under optimized LC conditions (pH 10, 20 mM ammonium formate) specifically chosen to improve the loadability of high-abundance peptides [22]. Considerations have likewise been made to ensure appropriate loadability in the second dimension. It is critical for the second dimension column to exhibit high peak capacity under high-mass load conditions, even when operated with the low pH/low ionic strength buffers preferred for positive ion electrospray ionization. Although the sample loading requirements for the second chromatographic dimension are not as stringent as the first dimension, as the column only needs to accommodate a fraction of the entire sample for each second dimension separation (e.g. 20-50 µg), the resolving capability of the second dimension separation under relatively high loading conditions is absolutely critical for the identification of low abundance HCP peptides. Interestingly, it was recently reported that a novel charged surface hybrid (CSH) stationary phase exhibits ideal chromatographic properties for this application. This stationary phase contains a low-level amount of basic moieties bonded to an organosilica, ethylene-bridged hybrid (BEH) C18 fully porous 1.7 µm particle [23]. Under acidic conditions, the CSH C18 stationary phase is positively charged and is able to significantly reduce the secondary interactions of charged peptides. The CSH C18 columns have therefore demonstrated improved chromatographic performance under high-mass load conditions [23], and they are a critical component of the HCP

8 ACS Paragon Plus Environment

Page 8 of 24

Page 9 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

application because they allow up to 5-fold increase in the amount of sample loaded for each assay (200-1,000 µg of digested mAb peptides loaded onto the first dimension column). The chromatographic performance of a microscale CSH C18 1.7 µm column (0.3 mm x 150 mm) has been compared with the performance of a column packed with similarly sized, sub-2 µm silica-based C18 particles (1.8 µm). The same amount of sample (200µg of NIST mAb digest) was loaded under basic conditions (pH 10) on the first dimension column of the 2D-LC RP/RP setup shown in Figure 1. The sample was fractionated by RP chromatography using a 10step elution protocol with increasing acetonitrile (ACN) concentrations (see the LC conditions in the Experimental section for details). The base-peak ion chromatograms for an example set of fractions (3rd, 5th, 6th, 9th and 10th fraction) with the CSH C18 versus silica-based C18 stationary phase are displayed in the Supplemental Figure S1 (panels A-E). Chromatograms from separations with the CSH C18 phase were found to consistently show improved (narrower) peak shapes for the abundant mAb peptides. This phenomenon is likely to be due to the fact that under acidic conditions (0.1% FA) the CSH stationary phase is positively charged and is able to significantly reduce the undesired interactions of high-abundance charged peptides (either solute-solute, solute-sorbent, or a combination of both) that have been hypothesized to be responsible for poor peak shapes [23]. In addition to improved peak capacity, it has been found that the CSH C18 phase, with its positive surface potential displays a different selectivity and retentivity when compared to the conventional silica-based C18 phase. In particular, peptides have been found to elute in 2-4% lower concentrations of ACN; the more basic residues a peptide contains, generally the more significant the change in selectivity between CSH C18 and conventional C18. This is likely a

9 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 24

consequence of electrostatic repulsion between the positive surface potential of the CSH column and the positive charges that a peptide bears under acidic RP chromatographic conditions. The superiority of the CSH column over the T3 column in terms of improved peak capacity under overloading conditions traslates into more HCP identifications as demonstrated by the results presented in Table IS (Supplemental Materials). In a separate experiment, an attempt was made to load the same amount of sample (200µg of NIST mAb digest) on a longer CSH column (0.3 mm x 250 mm) packed with 1.7 µm CSH particles. The sample was loaded directly (without using the 2D-LC setup) to evaluate the loading capacity of this CSH column under acidic conditions (0.1% FA). The experiment failed, as peptide breakthrough was starting to occur during sample loading (data not shown), even though this alternative CSH column was longer than the CSH columns used for the 2D-LC setup (250 vs 150 mm) and theoretically had a higher sample capacity. It is highly likely that the non-optimized conditions (low pH, low ionic strength mobile phase) used for loading the digest containing many high-abundance charged peptides were responsible for this outcome. This result clearly shows that a single chromatographic dimension does not have the ability to cope with the large discrepancy in concentration (4-6 orders of magnitude) between the mAb peptides and the HCP peptides. The advantage offered by the increased sample loading capacity of the CSH column is best exploited for improved HCP assays, only if is used in conjunction with further modifications of the original HCP assay. The increase in the amount of peptides loaded on the CSH column challenges the mass spectrometric ion detection, as additional isotopic distributions, populating the same m/z space, need to be resolved. This challenge is often referred as mass spectral crowding [24]. To address this challenge, we added an extra dimension of separation by

10 ACS Paragon Plus Environment

Page 11 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

employing a travelling-wave ion mobility (IM) cell [25] for a gas phase separation of co-eluting peptide precursors. The advantages offered by IM separations for detection of low-abundance peptides in complex samples have been previously demonstrated [26]. The IM separation hada two-fold effect on mass spectrometric ion detection: (i) it simplified the complexity of MS spectra such that the detection of low abundance precursors was enhanced; (ii) it improved the alignment of peptide precursors with their corresponding fragments following data independent acquisition (here referred as HDMSE). Figure 2 illustrates the utility of ion mobility (IM) separation to reduce spectral complexity at the precursor level. The top panel (2A) displays the combined electrospray ionization (ESI)-MS spectrum (10 scans) recorded during the elution of an HCP peptide from NSFL1 protein (see Table I), while the bottom panel (2B) shows the composite spectrum obtained after combining 7 ion mobility bins containing the same peptide precursor. Following the IM separation, the ions of interest from NSFL1 HCP protein (peptide SYQDPSNAQFLESIR) can be separated away from many other co-eluting peptide ions. These panels clearly demonstrate that the number of co-eluting precursors left after IM separation is significantly reduced. The incorporation of IM separations has also reduced the complexity of fragmentation spectra. Following the IM separation, peptide precursors are fragmented to produce more complex fragmentation spectra. Because high-energy fragmentation (HDMSE) takes place post IM separation, it is expected that peptide precursors and their corresponding fragments would align on the scale of IM drift times. Figure 2C-D depicts this observation clearly, as the apex of the extracted ion mobilogram for HCP peptide (m/z=877.92/+2 vs drift time) aligns well with the apex of the extracted ion mobilogram that has been recorded for one of the most abundant fragment ions predicted for that precursor sequence (m/z=1261.65, y11). This correlation

11 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 24

between the drift times of precursors and their fragments in conjunction with retention time correlation is used by data processing software, Protein Lynx Global Server (PLGS), to filter data into cleaner (complexity-reduced) fragmentation spectra, which greatly enhances the ability of the PLGS algorithm to identify low abundance HCPs [27]. Since IM separations are taking place in the gas phase, in a nitrogen filled travelling wave cell situated remotely from the ion source, the drift times of precursors and fragments originating from low-abundance HCPs are not affected by the presence of other high-abundance ions. The ion mobilograms presented in panels C and D from Figure 2 were recorded using data independent acquisition (HDMSE). Note that drift times for the precursor ion of HCP peptide SYQDPSNAQFLESIR and its y11 fragment are very similar, even though the IM cell contained a high excess of "matrix" ions introduced from the LC separation. On this basis, the reproducibility of IM separations, regardless of the presence or absence of high-abundance ions, has been used to accurately measure the collision cross section (CCS) for a variety of ions derived from lipids, drug metabolites, glycans [29-32] and peptides [33]. Traditionally, the collision energy applied in the transfer cell of the SYNAPT G2-S instrument is ramped over a relatively wide energy range (15 to 45 V) to achieve the fragmentation of multiply protonated peptide precursors (1-6 charge states) distributed over a wide m/z range. As a result, it is rather challenging to define the optimum collision energy for each individual precursor during the LC-MS run. Tenzer and coworkers have shown that the incorporation of an IM separation can be utilized to improve the fragmentation efficiency of peptides if the applied collision energy is correlated with the ion mobility drift time of peptide precursors [28]. Instead of applying a generically wide CE ramp (~30 V) for peptides eluted from a chromatographic separation at a given moment, the recently developed data acquisition method applies a much

12 ACS Paragon Plus Environment

Page 13 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

narrower, quasi "fixed", optimized CE ramp (~2 V) to a subset of peptide ion populations which are isolated by IM. In the original work describing this method, fragmentation efficiency was claimed to be improved by approximately 2 fold. The fragmentation spectra displayed in the Supplemental Figure S2 show a direct comparison between the wide ramp CE fragmentation spectra and the narrow ramp fragmentation spectra as recorded for a low-abundance peptide in a complex peptide sample. The intensity of the major fragment ions of this peptide (ENL T43) nearly doubled in the spectrum acquired under optimized CE conditions, confirming that this type of fragmentation achieves higher efficiency. The utility of data-independent HDMSE acquisition to correctly identify low-abundance HCPs is illustrated by the spectra shown in Figure 3 (panels A-C). The high-energy fragmentation spectra shown in these panels were acquired for the HCP peptide SYQDPSNAQFLESIR from NSFL1 protein. Each panel displays a portion of the fragmentation spectrum obtained without the IM separation (top spectrum), with precursor level IM separation (middle spectrum) and the confirmatory ESI-IM-MS/MS spectrum (shown at the bottom). The full range high-energy fragmentation spectra are displayed in Figure S3 in the Supplemental Material. The bottom spectra were recorded for a quadrupole selected (~2 Da mass window), ion mobility separated peptide precursor (m/z=877.92/+2). Three major fragment ions of this HCP peptide are clearly distinguished in all the HDMSE spectra (see the middle spectra in these panels). In the absence of the ion mobility separation (top spectra in each panel), the MS/MS background is significantly more complex and the spectra have more interfering fragment ion signals, thus preventing the software to make correct sequence identifications. Ion mobility separation at the precursor level enables the removal of a significant amount of interfering signals (including overlapping fragment ions and noise), resulting in less crowded, more discernible high energy fragmentation

13 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 24

spectra, which help to produce correct sequence identifications... It should be kept in mind that HDMSE acquisition is particularly useful because unlike traditional quadrupole-based precursor selection, it can employ very fast scan rates for mobility isolated sets of multiple precursor ions. The application of HDMSE acquisition to an HCP assay is therefore noteworthy, as it improves the duty cycle of the analysis and yields more in-depth sample coverage. The combination of the aforementioned developments has provided remarkable improvements in the detection limits and the utility of the 2D-RP/RP LC-MS strategy for HCP analysis. In demonstration of this, we have applied our methodologies in an inter-laboratory study for the analysis of HCPs in a candidate mAb reference material from NIST. The NIST mAb was digested and analyzed in triplicate (500 µg digest per injection) by three independent laboratories, located at different sites, using the 10-step 2D-LC high-pH RP/low-pH RP fractionation scheme described in the Experimental section. Four unique protein digest standards (ADH, PHO, BSA and ENL), originating from species other than the murine cells of the host were spiked into the NIST mAb digest post-digestion. These protein standards were used as internal calibrants to probe the dynamic range of the assay and to provide an internal reference for quantification of HCPs using the summed signal of the three best responding peptides of each protein identified in the analysis [34-35]. The results of the 2D-LC/HDMSE analysis of the NIST mAb are summarized in the Venn diagram displayed in Supplemental Figure S4. A total of 34 HCPs were found after combining individual laboratory results, with each laboratory achieving the identification of HCPs down to single digit ppm levels. Notably, 27 of theses HCPs were identified by at least 2 sites, and 14 out of the 34 HCPs were identified by all three sites. Moreover, each laboratory successfully identified each of the four spiked protein standards with concentrations ranging from 5-200 ppm (ADH, PHO, BSA). Individual

14 ACS Paragon Plus Environment

Page 15 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

HCP amounts (expressed in femtomoles) were estimated using the signals of the three best responding peptides from each protein compared against the signal of the top three peptides from one of the protein calibrants - PHO, loaded at 1,000 fmoles on-column (so called Hi3 method [34-35]). Based on these calculated molar amounts, the ng amounts of each HCP loaded oncolumn were calculated after taking into account the predicted molecular weight of each identified protein. HCP concentrations expressed in ng/mL were then computed based on the injected sample volume (250 µl). Finally, using the initial mAb concentration (100 mg/mL), the individual HCP concentrations (expressed in ppm, or ng HCP/mg mAb) were obtained. Each individual HCP concentration was calculated by each laboratory from three replicate injections. Table IIS (Supplemental Materials ) provides a detailed list of all the peptides belonging to 12 low-abundance (1-20 ppm) HCPs identified and quantified in this study. Following HCP identification, eleven out of twelve low-abundance HCPs (1-20 ppm) were subsequently validated by acquiring targeted "confirmatory" ESI-IM-MS/MS fragmentation spectra: targeted peptide precursors were isolated using the quadrupole mass filter (~2 Da mass window), separated from other co-eluting isobaric precursors using ion mobility, and fragmented with an optimized (fixed) CE in the transfer cell of the IM-enabled mass spectrometer. These ESI-IMMS/MS spectra are presented in Figure S5, while Figure 4 displays the fragmentation spectra recorded for three of the very low-abundance HCPs (in the 2-7 ppm range). These "cleaner", information rich, high-quality MS/MS spectra contain extensive amino acid sequence coverage for each peptide, thus providing high confidence validation to the HCP identifications made from 2D-LC/HDMSE experiments. The newly developed HCP assay was also used to identify and quantify low-abundance HCPs in formulations of Infliximab (trade name Remicade), a licensed biopharmaceutical mAb, used for the treatment of several diseases including Crohn's disease and

15 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 24

rheumatoid arthritis. The 2D-LC/HDMSE analysis revealed the presence of two HCPs in the concentration range of 10-25 ppm: epidermal growth factor-like protein 8 (25 ppm) and WD repeat-containing protein 37 (15 ppm). These HCP findings were also confirmed (validated) by the ESI-IM-MS/MS spectra of two selected peptides displayed in Figure S6. It is worth noting that even if the identified peptides contain only very short amino acid sequences (7 AA long), they are very unique. The uniqueness of these sequences was confirmed with a Protein BLAST search (http://blast.ncbi.nlm.nih.gov/Blast.cgi) against the 16,648 protein sequences from the mouse proteome (UniProt).

Conclusions

An improved HCP assay based on comprehensive 2D-LC separations (high pH/low pH RP/RP) coupled to ion mobility high-resolution (~20,000) QTOF mass spectrometry detection was successfully used for identification and quantification of low-abundance HCPs from high-purity mAbs. A newly developed stationary phase containing a modified hybrid particle technology (CSH - charge surface hybrid) was utilized to enhance the sample loadability of the LC system. Ion mobility separations of peptide precursors further improved the separation of complex peptide mixtures, and enhanced the quality of high-energy fragmentation spectra thus allowing the HCP assay to achieve previously unattainable detection limits (1 ppm HCP, or 1 ng HCP/mg mAb). The results of the 2D-LC/HDMSE assay were validated by acquiring "confirmatory" MS/MS spectra from quadrupole isolated (~2Da isolation window), mobility separated peptide precursors. Most importantly, the performance of this methodology in the inter-laboratory study demonstrates that this HCP assay exhibits noteworthy reproducibility.

16 ACS Paragon Plus Environment

Page 17 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Furthermore, these results indicate that mass spectrometric-based assays are now able to attain comparable sensitivity to traditional HCP assays (e.g. ELISAs), while offering the unique advantage of providing unambiguous HCP identification.

Acknowledgements

The authors would like to thank Karen Haas and Leslie Malouin

from Waters Corp. for

preparing Figure 1 of the manuscript.

Supporting Information Available

Full details regarding the experimental conditions used for sample preparation and for system operation are presented in the Supplementary Materials section. This section also includes six supplemental figures (S1-S6) and two supplemental tables (I-IIS) along with enhanced versions of all manuscript figures and tables. This information is available free of charge via the Internet at http://pubs.acs.org/.

Figure Captions Figure 1. Fluidic configuration for the two-dimensional high pH reversed phase/low pH reversed phase (RP/RP) chromatographic setup employing on-line dilution. Each dimension of the RP/RP chromatographic separation is individually controlled by the binary solvent modules BMS1 and BMS2, while the auxiliary solvent manager (ASM) is responsible for providing the make-up flow for on-line dilution. Figure 2. The role of ion mobility (IM) separation in reducing spectral complexity: (A) ESI-MS spectrum generated from 10 combined scans recorded during elution of an HCP peptide (SYQDPSNAQFLESIR) without IM separation; (B) ESI-IM-MS spectrum recorded for the same peptide obtained after combining 7 mobility bins; (C) mobilogram recorded for the precursor of this peptide (m/z=877.92/+2); (D) mobilogram obtained for one of the most abundant fragments (y11 = 1261.65) of the peptide. 17 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 24

Figure 3. Comparison of high-energy fragmentation spectra obtained with/without ion mobility separation. Each panel displays a portion of the fragmentation spectrum obtained without the IM separation (top spectrum), with precursor level IM separation (middle spectrum) and the confirmatory ESI-IM-MS/MS spectrum (shown at the bottom). The bottom spectra were recorded for a quadrupole selected (~2 Da mass window), ion mobility separated peptide precursor (m/z=877.92/+2). All spectra are centered around three abundant fragment ions of this HCP peptide: (A) b4; (B) y11(+2) and (C) y11(+1). Figure 4A-C. HCP validation of proteins identified in the NIST mAb by the 2D-LC/USMSE approach: "confirmatory" ESI-IM-MS/MS spectra recorded for three peptides from 3 lowabundance HCPs (2-7 ppm range). Extensive peptide sequence coverage was obtained in each case, with almost complete sequence verification. Table I. Fourteen HCPs, present in the concentration range of 1-100 ppm, were identified in the NIST mAb by all three laboratories. Eleven out of twelve low-abundance HCPs (1-30 ppm, were validated using "confirmatory" MS/MS spectra. Three confirmatory MS/MS spectra displayed in Figure 4A-C were acquired for the three HCPs labeled by the blue arrows shown this table.

References

1. Doneanu C, Xenopoulos A, Fadgen K, Murphy J, Skilton SJ, Prentice H, Stapels M, Chen W mAbs 2012, 19, 242. 2. Kabir A, Hamlet C, Malik A J Chromatogr A 2004, 1047, 1. 3. Stanley JS, Sack TM, Tondeur Y, Beckert WF Biomed Environ Mass Spectrom 1998, 17, 27. 4. ICH Harmonised Tripartite Guideline. Specifications: Test Procedures and Acceptance Criteria for Biotechnological/Biological Products Q6B. 1999. 5. Schenauer MR, Flynn GC, Goetze AM, Anal Biochem 2012, 428, 150. 6. Schenauer MR, Flynn GC, Goetze AM, Biotechnol Prog 2013, 29, 951. 7. Tscheliessnig AL, Konrath J, Bates R, Jungbauer A Biotechnol J 2013, 8, 655. 8. Zhang Q, Goetze AM, Cui H, Willie J, Trimble S, Hewig A, Flynn GC, mAbs 2014 6, 659. 9. Thompson JH, Chung WK, Zhu M, Tie L, Lu Y, Aboulaich N, Strouse R, Mo WD Rapid Commun Mass Spec 2014, 28, 855. 10. Aboulaich N, Chung WK, Thompson JH, Larkin C, Robbins D, Zhu M Biotechnol Prog 2014, 30, 1114.

18 ACS Paragon Plus Environment

Page 19 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

11. Rogers R, Bell A, Kowski T, Bailey R, presentation at the 61st ASMS Conference on Mass Spectrometry and Allied Topics, 2013, Minneapolis, MN. 12. Yu C, presentation at the 10th Symposium on the Practical Applications of Mass Spectrometry in the Biotechnology Industry, 2013, Boston, MA. 13. Joucla G, Le Senechal C, Begorre M, Garbay B, Santarelli X, Cabanne C J Chromatogr B 2013, 942-943, 126. 14. Reisinger V, Toll H, Mayer RE, Visser J, Wolschin F Anal Biochem 2014, 463, 1. 15. Xiao G, Da Ren, Bondarenko P, presentation at the 62nt ASMS Conference on Mass Spectrometry and Allied Topics, 2014, Baltimore, MD. 16. Sperry JB, presentation at the First Annual BEBPA Workshop on Host Cell Proteins Assays, 2014, May 16, Dubrovnik, Croatia. 17. Wang X, Hunter AK, Mozier NM Biotech& Bioeng 2009, 103, 446. 18. Bomans et al. Plos One 2013, 8, 1. 19. Zhu-Shimoni et al. Biotechnol & Bioeng 2014, 111, 2367. 20. Gritti F, Guiochon G, J Chromatogr A 2005, 1099, 1. 21. McCalley DV Anal Chem 2006, 78, 2532. 22. Gritti F, Guiochon G, J Chromatogr A 2009, 1216, 63. 23. Lauber MA, Koza SM, McCall SA, Alden BA, Iraneta PC, Fountain KJ Anal Chem 2013, 85, 6936. 24. Michalski A, Cox J, Mann M J Proteome Res 2011, 10, 1785. 25. Giles K, Pringle SD, Worthington KR, Little D, Wildgoose JL, Bateman RH Rapid Commun Mass Spectrom 2004, 18, 2401. 26. Baker E et. al. J Proteome Res 2010, 9, 997. 27. Geromanos SJ, Hughes C, Ciavarini S, Vissers JPC, Langridge JI, Anal Bioanal Chem 2012, 404, 1127. 28. Distler U, Kuharev J, Navarro P, Levin Y, Schild H, Tenzer S, Nat Meth 2014, 11, 167. 29. Cuyckens F, Wassvik C, Mortishire-Smith R, Tresadern G, Campuzano I, Claereboudt J Rapid Commun Mass Spectrom 2011, 25, 3497. 30. Pagel K, H30. Pagel K, Harvey DJ Anal Chem 2013, 85, 5138.

19 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 24

31. Paglia G, Williams JP, Menikarachchi L, Thompson JW, Tyldesley-Worster R, Halldorsson S, Rolfsson O, Moseley A, Grant D, Langridge J, Palsson BO, Astarita G Anal Chem 2014, 86, 3985. 32. Paglia G, Angel P, Williams JP, Richardson K, Olivos HJ, Thompson JW, Menikarachchi L, Steven Lai, Callee Walsh, Moseley A, Plumb RS, Grant D, Palsson BO, Langridge J, Geromanos S, Astarita G Anal Chem 2015, 87, 1137 33. Lietz CB, Yu Q, Li L J ASMS 2014, 25, 2009. 34. Silva JC, Gorenstein M, Li GJ, Vissers P, Geromanos SJ Moll Cell Proteomics 2006, 1, 144. 35. Silva JC, Denny R, Dorschel C, Gorenstein M, Li GJ, Richardson K, Wahl D, Geromanos SJ Moll Cell Proteomics 2006, 5, 589.

20 ACS Paragon Plus Environment

Page 21 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 1

Figure 2

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3

ACS Paragon Plus Environment

Page 22 of 24

Page 23 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 4

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Table I Note for Reviewers: The full (enlarged) versions of the Figures 1-4 and Table I that are included in the manuscript, are available in the Supplementary materials

ACS Paragon Plus Environment

Page 24 of 24