Development of Comprehensive Online Two ... - ACS Publications

Apr 4, 2018 - Agilent Technologies, Santa Clara, California 95051, United States. §. School of Pharmaceutical Sciences, University of Geneva and Univ...
1 downloads 0 Views 990KB Size
Subscriber access provided by UNIV OF CAMBRIDGE

Development of Comprehensive Online Two-Dimensional Liquid Chromatography-Mass Spectrometry using Hydrophilic Interaction and Reversed-Phase Separations for Rapid and Deep Profiling of Therapeutic Antibodies Dwight R Stoll, David Christopher Harmes, Gregory O. Staples, Oscar G. Potter, Carston T. Dammann, Davy Guillarme, and Alain Beck Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b00776 • Publication Date (Web): 04 Apr 2018 Downloaded from http://pubs.acs.org on April 4, 2018

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 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 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.

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 8 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

Development of Comprehensive Online Two-Dimensional Liquid Chromatography-Mass Spectrometry using Hydrophilic Interaction and Reversed-Phase Separations for Rapid and Deep Profiling of Therapeutic Antibodies Dwight R. Stoll*,1, David C. Harmes1, Gregory O. Staples2, Oscar G. Potter2, Carston T. Dammann1, Davy Guillarme3, and Alain Beck4 1) Department of Chemistry, Gustavus Adolphus College, St. Peter, Minnesota 56082, United States 2) Agilent Technologies, Santa Clara, CA, USA 3) School of Pharmaceutical Sciences University of Geneva, University of Lausanne, Rue Michel Servet 1, 1206 Geneva 4, Switzerland 4) Institut de Recherche Pierre Fabre, Center of Immunology Pierre Fabre, 5, Avenue Napoléon III, BP 60497, 74160 SaintJulien-en-Genevois, France ABSTRACT: Monoclonal antibodies (mAb) and related molecules are being developed at a remarkable pace as new therapeutics for the treatment of diseases ranging from cancer to inflammatory disorders. However, characterization of these molecules at all stages of development and manufacturing presents tremendous challenges to existing analytical technologies because of their large size (ca. 150 kDa) and inherent heterogeneity resulting from complex glycosylation patterns and other post-translational modifications. Multi-dimensional liquid chromatography is emerging as a powerful platform technology that can be used to both improve analysis speed for these molecules by combining existing one-dimensional separations into a single method (e.g., Protein A affinity separation and size-exclusion chromatography), and increasing the resolving power of separations by moving from one dimension of separation to two. In the current study, we have demonstrated the ability to combine hydrophilic interaction (HILIC) and RP separations in an online comprehensive 2D separation coupled with high resolution MS detection (HILIC×RP-HRMS). We find that Active Solvent Modulation (ASM) is critical for coupling these two separation modes because it mitigates the otherwise serious negative impact of the acetonitrile-rich HILIC mobile phase on the second dimension RP separation. The chromatograms obtained from these HILIC×RP-HRMS separations of mAbs at the subunit level reveal the extent of glycosylation on the Fc/2 and Fd subunits in analysis times on the order of two hours. In comparison to previous CEX×RP separations of the same molecules, we find that chromatograms from the HILIC×RP separations are richer and reveal separation of some glycoforms that co-elute in the CEX×RP separations.

Monoclonal antibodies (mAbs) and their related compounds are a major class of therapeutics for the treatment of cancer, inflammatory and autoimmune disorders, and infectious diseases. To date, the FDA and EMA have already approved more than 50 therapeutic mAbs and there are over 500 candidates in clinical studies.1 In the near-term, it is expected that 10 to 12 mAbs will enter the market each year.2 The success of mAbs is related to their specificity and their various mechanisms of action, leading to an improvement of the benefit-risk ratio for patients, compared to other treatments.3, 4 Based on current forecasts for 2018, there are six mAbs among the top 10 selling drugs, and adalimumab, at the top of the list, is expected to generate almost $20 billion revenue in one single year.5 However, characterization of antibodies is challenging because of their large size (ca. 150 kDa) and complex posttranslational modifications including extensive glycosylation.6 The analytical challenges associated with characterizing mAbs

and related materials have led many groups to develop new chromatographic and electrophoretic methods, and many of them have been coupled with mass spectrometric (MS) detection for direct determination of protein mass, to enable primary structure assessment, and for determination of posttranslational modifications.7 Hydrophilic interaction chromatography (HILIC) is particularly attractive for the characterization of glycosylated proteins because of the exquisite selectivity of this separation mode for glycans. Indeed, HILIC separations have also been coupled to MS detection for rapid profiling of mAb glycoforms.8–11 In our own work we have observed that one major limitation of existing column materials is that they require a significant level of trifluoroacetic acid (TFA; ca. 0.1% (v/v)) in the mobile phase to obtain good retention and peak shape for proteins in the HILIC mode. This level of TFA suppresses ionization of the proteins during electrospray ionization mass spectrometry (ESI-MS), leading to poorer sensitivity compared to what is observed with re-

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

versed-phase separations followed by ESI-MS using formic acid as a mobile phase modifier.10 Over the past couple of years, there has been intense interest in the biopharmaceutical community in developing twodimensional liquid chromatography (2D-LC) methods to improve the efficiency and/or depth of mAb analyses.12, 13, 6, 14 For example, Protein A affinity separations have been combined with size-exclusion chromatography (SEC) in a 2D-LC format for efficient determination of mAb titer and aggregation in cell culture supernatant.15, 16 Many of these applications have been based on heart-cutting strategies to focus on one or several constituents of a sample as they elute from a first dimension (1D) separation. 17, 16 In our own work, we have demonstrated the utility of comprehensive on-line 2D-LC separations involving cation-exchange (CEX) and reversedphase (RP) separations with MS detection for deep profiling and facile comparison of mAb originators and biosimilars.18 One attractive option for 2D-LC separation of intact or partially digested mAbs is the coupling of HILIC and RP separations. This approach has the potential to not only improve upon the separation of protein glycoforms provided by the HILIC separation through addition of the second dimension (2D) RP separation, but also improve the sensitivity of ESI-MS detection for these molecules because the analytes can be presented to the MS in formic acid containing mobile phase rather than the TFA-containing mobile phase required for HILIC separations. Such separations have been demonstrated for peptides and a variety of small molecule applications.19, 20 Experimentally, this combination of separations has been challenging because the high acetonitrile (ACN) content of HILIC mobile phases (typically greater than 60%) produces broadening and distortion of peaks in the second dimension (2D) separation, especially when large volumes of first dimension (1D) effluent are injected into the 2D column.21–23 This challenge is usually met by decreasing the volume of 1D effluent that is injected into the second dimension, increasing the 2 D column volume, or both.19, 24, 25 However, this limits the detection sensitivity of such separations because of the dilution of analytes during the 2D separation step.26, 22 An alternative to simply reducing the volume of 1D effluent injected into the 2D column is to dilute the fraction of 1D effluent with ‘weak solvent’ during the transfer from the first to the second dimension column. When RP separation is used in the second dimension, this means water is added as the diluent. The benefit of this approach was first demonstrated decades ago by Oda and coworkers27, and has been implemented in numerous ways in 2D-LC separations since then.28 Recently, we introduced a valve-based implementation of this approach we refer to as Active Solvent Modulation (ASM), and demonstrated the benefit to 2D-LC separations of peptides when RP separations were used in both dimensions.23 In the current work, we have demonstrated the benefit of ASM when coupling HILIC and RP separations for comprehensive 2DLC-MS analysis of mAbs. The comprehensive approach is the most efficient way to assemble a complete list of all detectable species in a complex mAb material. We have judiciously chosen three antibodies that differ substantially in their degree of N-glycosylation ranging from aglycosylated (atezoluzumab) to glycosylation only on the Fc subunit (obinutuzumab, glyco-engineered), to glycosylation on both the Fc and Fd subunits (cetuximab). This series of molecules enables a clear illustration of the benefits of both HILIC and

RP separations combined in a 2D separation format, followed by MS detection. EXPERIMENTAL SECTION MATERIALS. All reagents and materials were used as obtained from their respective manufacturers: acetonitrile (ACN, Chromasolv LC-MS grade), formic acid (FA; part no. 09676), trifluoracetic acid (TFA; part no. T6508), Amicon Ultra 0.5 mL centrifugal filters (10 kDa device, part no. Z67710896EA) and DTT (part no. 43815) were all from Sigma Aldrich (St. Louis, MO). Water was purified in-house using a Milli-Q water purification system (Billerica, MA). The IdeS protease enzyme was obtained from Genovis AB (part no. A0-FRI-008; Lund, Sweden).29 Cetuximab (Erbitux®, Eli Lilly and Company), obinutuzumab (Gazyvaro®, Roche), and atezolizumab (Tecenriq®, Roche), were obtained from their respective manufacturers. SAMPLE PREPARATION. Partial digestion of intact mAbs was carried out by addition of 100 µg of mAb to a polypropylene tube containing 100 units of lyophilized IdeS enzyme, and the total volume was brought to 70 µL using IdeS digestion buffer (50 mM sodium phosphate, 150 mM sodium chloride, pH 6.6). This solution was incubated at 37°C for 45 min. In some cases, as indicated, this digested mAb solution was concentrated prior to analysis using an Amicon Ultra 10 kDa filtration device, according to the manufacturer’s protocol. Reduction of interchain disulfide bonds was carried out by transferring an aliquot of the IdeS digest solution to a polypropylene tube and the addition of 10% (v/v) of 1M DTT in water. The solution was incubated at 37°C for 30 min and then analyzed immediately. INSTRUMENTATION. All instrument modules were from the 1290 Infinity line from Agilent Technologies (Waldbronn, Germany): first (1D) dimension binary pump (Model G4220A); second (2D) dimension high speed pump (Model G7120A); autosampler (Model G4226A), thermostated column compartments (Model G1316C), and diode-array UV detectors (DADs) (Model G4212A and Model G7117A). The interface valve connecting the two dimensions of the system (p/n: 5067-4266) was set up with two nominally identical 5, 20 or 40 µL sample loop sets. A pressure release kit (p/n: G4236-60010) was installed between the outlet of the 1D detector and the inlet to the interface valve to minimize disturbances in the 1D detector baseline. A pressure relief valve (p/n: G4212-60022) was used between the 2D DAD and the mass spectrometer nebulizer. The mass spectrometer was a time-of-flight (TOF-MS) instrument (Agilent model G6230B) equipped with an Agilent JetStream electrospray ionization source. The mass analyzer was calibrated using a standard tuning compound mixture (Agilent, part no. G1969-85000). Hexakis (1H,1H,3Hperfluoropropoxy) phosphazene was used as a reference mass (m/z 922.0098) compound for calibration of mass spectra prior to deconvolution, sprayed continuously into the JetStream source from a secondary reference nebulizer. The 2D-LC instrument was controlled by OpenLab Chromatography Data System (C.01.07), with a 2D-LC Add-on (rev. A.01.04 [017]). First dimension HILIC separations were performed using two 150 mm x 2.1 mm i.d. Acquity UPLC Glycoprotein Amide (300Å, 1.7 µm; Waters Corporation, Milford MA) columns connected in series. Gradient elution was used in the first dimension, where solvent A contained

ACS Paragon Plus Environment

Page 2 of 8

Page 3 of 8 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 0.3% (v/v) trifluoroacetic acid in water and solvent B was ACN. The HILIC gradient used for each mAb varied slightly as follows: cetuximab - 69-65.5-65.5-61-61-69-69 %B from 010-70-112.5-120-120.01-150 min; obinutuzumab - 69-66-6660-69-60-60 %B from 0-20-80-90-95-95.01-125 min; atezolizumab - 71-69-69-60-60-71-71 %B from 0-2-70-80-90-90.01125 min. The column temperature was 40°C, and the flow rate was 0.040 mL/min. Unless otherwise indicated, 4 µL of sample was injected using an injection program that involved drawing 8 µL of ACN into the injector needle after drawing up the analytical sample to mitigate the effects of the water rich sample on the performance of the HILIC separation.30 In cases where less than 4 µL of sample was injected, the ratio of sample to ACN remained 1:2. Second dimension reversed-phase separations were performed using a 30 mm x 2.1 mm i.d. PLRP-S (1000 Å column, 5 µm; Agilent Technologies). Solvent A was 0.1% (v/v) formic acid in water, and solvent B was ACN. The column temperature was 75°C, the flow rate 1.0 mL/min. In each 2D separation cycle 40 µL of 1D column effluent was diluted 1:2 (1 part effluent, 2 parts 2D mobile phase) using the inline dilution feature of the ASM interface valve, and the modulation time was 60 s. The 2D gradient elution program from 2-2-25-37-50-2-2 %B from 0-0.41-0.420.84-0.92-0.93-1.0 min, where the hold at 2 %B for the first 24 s is required for inline dilution of the injected sample. Following detection, the effluent from the 2D column was split 1:1 prior to the mass spectrometer using a simple T-piece. UV absorbance data were collected at 280 nm. In the first dimension, the acquisition rate was 2.5 Hz; in the second dimension, the acquisition rate was 80 Hz. The 2D chromatograms presented as contour plots in Figs. 1-5 were produced using LC Image (GC Image, Lincoln, NE) version 2.7, release 2. TOF-MS data were acquired from m/z 500 to m/z 6000 in positive ion mode at an acquisition rate of 5 Hz. The drying gas temperature and flow rate were 325°C and 10 L/min, respectively, while the sheath gas temperature and flow rate were 350°C and 11 L/min, respectively. The nebulizer gas pressure was 35 psi. The capillary and nozzle voltages were 5.5 and 1.0 kV, respectively, and the fragmentor voltages set to 300 and 350 V, unless otherwise specified. Parameters used for MaxEnt deconvolution of protein mass spectra were as follows: m/z range 15,000–35,000 for Fd, Fc/2, and LC; step size 0.1 Da; top 25% of peak; Mass spectra were re-calibrated prior to deconvolution using a reference mass of m/z 922.0098. Throughout the text, glycoforms are designated using an abbreviated description of composition in the form HxNxFxNGNAx where H = hexose, N = N-acetyl. Glycoform identities listed in Tables S1-S3 were verified by comparison to the results of released glycan analysis by 1D-LC with fluorescence detection as shown in Fig. S5. RESULTS AND DISCUSSION METHOD DEVELOPMENT Data supporting choices of some important method parameters are provided as Supplemental Information. While these data are not the focus of the discussion here, readers interested in pursuing HILIC×RP separations will find the data informative. As mentioned above and touched on in previous work30, judicious selection of sample injection conditions is critical to enable injection of enough sample volume to detect low abundance proteoforms, without sacrificing chromatographic reso-

lution due to volume overload. The results of our experiments aimed at finding suitable injection conditions for this work are shown in Fig. S1. Similarly, the mass of protein injected affects detection sensitivity and can reduce chromatographic resolution due to mass overload. The results of experiments intended to understand the impact of the mass of mAb injected on 1D resolution under the conditions of the HILIC×RP separations are shown in Fig. S2. In the process of selecting flow rates for use in the first and second dimensions of these 2DLC separations, we also briefly examined the effect of flow rate on peak volume (and thus, resolution) for mAb subunit peaks in the first and second dimension separations. These results are shown in Figs. S3 and S4, respectively. ATEZOLIZUMAB The first dimension and 2D chromatograms for the HILIC×RP separation of the subunits of atezolizumab are shown in Fig. 1. In this case, the interpretation of the chromatograms is quite straightforward. Atezolizumab is aglycosylated (N298A mutation)31 and thus we only observe three major peaks corresponding to the Fc/2, LC, and Fd subunits of the antibody. In this case, we show the entire 2D time axis; in subsequent 2D chromatograms we only show the last 30-s of the 2D axis because approximately the first 20-s of each 2D cycle is spent injecting the large (effectively 100 µL after dilution by ASM) volume of 1D effluent into the 2D column. A list of all of the species detected for this antibody is provided as Supplementary Information in Table S1.

Figure 1. Separation of IdeS-digested and reduced atezolizumab by 1D-LC HILIC (A) and HILIC×RP (B). In this case 16 µg of mAb was injected, and detection was performed by UV absorption at 280 nm. The prominent peak at about 20 min in the first dimension and 5 s in the second dimension is dithiothreitol (DTT).

IMPACT OF ASM ON 2D SEPARATION QUALITY

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

The benefit of ASM to the HILIC×RP separations in this work is illustrated in Fig. 2. In this case, a shorter 1D separation (Fig. 2A) was used because the simplicity of Atz does not demand a longer separation, and because the focus of these experiments was investigating the impact of ASM on the 2D separation. Four different scenarios are compared; three where ASM is not used, and 5, 20, or 40 µL fractions of 1D effluent is injected directly into the 2D column, and one where ASM is activated and 40 µL of 1D effluent is injected into the 2D column after diluting 2.5:1 (i.e., 2.5 parts diluent and 1 part sample) with 0.1% FA from the 2D pump. Figure 2B shows that when ASM is not used, there is significant breakthrough of analytes in the dead volume of the 2D column, because the 1D effluent fraction contains much more ACN (typically about 65%) than the eluents used in the second dimension solvent gradient (25-37%). Even when the fraction volume is as small as 5 µL, there is a significant breakthrough peak. On the other hand, when ASM is activated and the fraction of 1D effluent is 40 µL prior to dilution, no breakthrough is observed by UV or

MS detection. Figure 2C shows a later portion of the 2D chromatograms focused on the elution window for the Fc/2 and Fd subunits. Because much of the protein breaks through in the column dead volume, detection sensitivity does not increase as expected with increasing fraction volume. In other words, the 2D peaks do not increase in height as the volume of 1 D effluent injected is increased. Activation of ASM provides the best detection sensitivity of the scenarios evaluated here, because all of the protein injected is first retained, and then eluted during the solvent gradient program. We attribute the faint horizontal bands (Figs. 2D-2F) observed at the dead time of the 2D column and again around 40 s to rapid changes in mobile phase composition during each 2D cycle. These occur upon the start of the injection step and then again when the mobile phase transitions from the 2% ACN used for ASM to the 25% ACN used at the point of the solvent gradient used for elution. Such rapid changes in solvent composition are known to produce apparent changes in absorbance when using UV detection.32

Figure 2. Effect of Active Solvent Modulation (ASM) on 2D separation quality. All chromatograms are based on detection by UV absorption at 280 nm. (A) Short 1D HILIC separation of IdeS digested and reduced atezolizumab; (B) Summed 2D chromatograms from 14-19 min. in the first dimension showing the breakthrough of protein that occurs at the 2D column dead time when ASM is not used. (C) Summed 2D chromatograms from 14-19 min. in the first dimension showing the dependence of detection sensitivity for the Fc/2 and Fd

ACS Paragon Plus Environment

Page 4 of 8

Page 5 of 8 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 subunits on loop volume and whether or not ASM is used. (D-G) 2D chromatograms for the four scenarios examined; the same intensity scales are used in all four contour plots. The acronym DTT refers to dithiothreitol.

OBINUTUZUMAB AND CETUXIMAB First dimension and 2D chromatograms for the HILIC×RP separations of IdeS-digested and reduced obinutuzumab and cetuximab are shown in Figures 3A/B and 3C/D, respectively. Whereas only three major peaks are observed in the separation of atezolizumab, many more major peaks are observed for both obinutuzumab and cetuximab. In the case of obinutuzumab, the Fc/2 subunit is glycosylated, resulting in a family of peaks eluting at the same 2D retention time, but nicely separated by the 1D HILIC column. This is highlighted by the black rectangle in Fig. 3B. However, since the Fd subunit of obinutuzumab is not glycosylated, we again only observe one major peak for that subunit in the chromatogram. A complete list of all of the species observed for obinutuzumab using HILIC×RP-MS is shown in Table S2.

Moving to the chromatograms for cetuximab we see that both the Fc/2 and Fd subunits are observed as families of peaks indicated by the black rectangles in Fig. 3D. This is because both the Fc/2 and Fd subunits of cetuximab are glycosylated. These separations clearly illustrate the value of the 2D RP separation in addition to the 1D HILIC separation that has excellent selectivity for the different glycoforms of each subunit. Although the 1D HILIC separation separates the glycoforms of a particular subunit effectively, several Fc/2 glycoform peaks are overlapped with Fd glycoform peaks along the 1 D separation axis (see, for example, peaks C.15 and C.31 in Fig. 3D). Representative mass spectra for high and low abundance species detected in the cetuximab sample are shown in Fig. S7.

Figure 3. 1D and HILIC x RP chromatograms for IdeS digested and reduced obinutuzumab (A/B) and cetuximab (C/D). In each case 16 µg of mAb was injected, and detection was performed by UV absorption at 280 nm. See Table S3 for identities of peaks C.15 and C.31.

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 4. Comparison of peak patterns for the proteoforms of IdeS-digested and reduced cetuximab separated by CEX × RP (adapted from 18) or HILIC × RP (B). Chromatograms are based on detection by UV absorption at 280 nm. In the case of the CEX × RP separation 5 µg of mAb digest was injected. Annotations are based on TOF-MS detection. See Tables S3 and S4 for identities of labelled peaks.

COMPARISON OF CEX×RP AND HILIC×RP SEPARATIONS In previous work, we used CEX×RP-MS separations to compare several mAbs and their counterpart biosimilar molecules.18 Figure 4 shows a comparison of the CEX×RP separation of IdeS-digested and reduced cetuximab to the HILIC×RP separation of the same sample. The improvement in resolution that the 1D column provides in the case of HILIC×RP over CEX×RP is quite striking. If we examine the elution patterns of species detected in both cases, we can rationalize this result. The single peak labelled Cx.4 in the CEX×RP chromatogram (Fig. 4A) contains several Fd species (modified by H5N4F1NGNA1, H6N4F1NGNA1, H7N5F1NGNA1, or H8N5F1NGNA1 glycans – please see Supporting Information for glycan nomenclature). Whereas all these peaks coelute in the case of the 1D CEX separation, they are well resolved by the 1D HILIC column into the peaks labelled C.33, 36, 39 and 42, as shown in Fig. 4B. A portion of the chromatogram shown in Fig. 4B is re-plotted in Fig. 5, focusing on the Fd subunit peak envelope (i.e., 40 to 130 min in the first dimension, and 46 to 56 s in the second dimension). The annotations of the major glycoforms detected across this envelope clearly indicate the increase in 1D retention time with increasing glycan size, and the power of the 1D HILIC separation for resolving these species. Separation of Fd subunits conjugated with different isomers of the same glycan is achieved; for example the H6N4F1 glycoforms eluting as distinct peaks around 80-85 minutes in the first dimension. This chromatographic separation prior to mass spectrometric detection should improve the quality of deconvoluted mass spectra for molecules like these, particularly in cases where there is a large difference in concentration of two species that would otherwise coelute in a conventional 1D separation (see for example the Fc and Fd species shows in Fig. S7 that coelute in the HILIC separation and are detected with a 10-fold difference in ion abundance).

Figure 5. Expanded view of the Fd subunit peak envelope of cetuximab from Fig. 4, with glycan structures indicated. Blue square – N-acetylglucosamine; White diamond – Nglycoylneuraminic acid; Green circle – Mannose; Yellow circle – Galactose; Red triangle – Fucose. CONCLUSIONS In this work we have demonstrated the ability to couple HILIC and RP separations in a comprehensive 2D-LC format coupled with high resolution mass spectrometric detection for the rapid and deep characterization of therapeutic antibodies. When using the HILIC separation in the first dimension, the high level of acetonitrile (ca. 65% (v/v)) in the mobile phase causes severe peak splitting and distortion when large fractions (> 5 µL) of the 1D effluent are injected into a 2D RP column for further separation. However, we have shown here that implementation of ASM to decrease the ACN content of these fractions prior to injection eliminates peak splitting, enabling injections of large fractions of 1D effluent into the 2D column, and thus sensitive detection of separated proteins.

ACS Paragon Plus Environment

Page 6 of 8

Page 7 of 8 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 Gstöttner, C. J.; Klemm, D.; Haberger, M.; Bathke, A.; Wegele, Using a series of three mAbs that vary greatly in extent of(7) H.; Bell, C. H.; Kopf, R. Analytical Chemistry. 2017, 90, 2119– N-glycosylation, we have shown that the chromatograms 2125. obtained from HILIC×RP-MS analyses rapidly reveal the extent of glycosylation on the Fc/2 and Fd subunits of each mAb. Furthermore, we find that the HILIC×RP separations are (8) D’Atri, V.; Dumont, E.; Vandenheede, I.; Guillarme, D.; Sandra, P.; Sandra, K. LCGC Europe. 2017, 30, 424–434. much more selective for separation of the numerous glycoforms of heavily glycosylated mAbs (e.g., cetuximab) compared to other types of 2D-LC such as CEX×RP. This is due (9) D’Atri, V.; Fekete, S.; Stoll, D.; Lauber, M.; Beck, A.; Guillarme, D. Journal of Chromatography B. 2018, 1080, 37– entirely to the high selectivity of the HILIC separation for 41. mAb subunits with different glycans. We believe the HILIC×RP-MS methods described here constitute a powerful addition to the rapidly developing toolkit for analysis of thera- (10) D’Atri, V.; Fekete, S.; Beck, A.; Lauber, M.; Guillarme, D. Analytical Chemistry. 2017, 89, 2086–2092. peutic antibodies based on multi-dimensional separations coupled with high resolution mass spectrometric detection. (11) Bobály, B.; D’Atri, V.; Beck, A.; Guillarme, D.; Fekete, S. Journal of Pharmaceutical and Biomedical Analysis. 2017, 145, 24–32.

ASSOCIATED CONTENT Supporting Information

(12) Wang, X.; Buckenmaier, S.; Stoll, D. Journal of Applied Bioanalysis. 2017, 3, 120–126.

A single document (PDF) containing additional data relevant to method development and detection protein species is provided as Supporting Information. This is available free of charge on the ACS Publications website.

(13) Stoll, D. R.; Maloney, T. D. LCGC North America. 2017, 35, 680–687.

AUTHOR INFORMATION

(14) Bathke, A.; Klemm, D.; Gstöttner, C.; Bell, C.; Kopf, R. LCGC Europe. 2018, 31, 10–21.

Corresponding Author * 800 West College Avenue, Saint Peter, MN 56082, [email protected].

Author Contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

ACKNOWLEDGMENT This work was supported by a grant from the United States National Science Foundation (CHE-1508159) and a Thought Leader Award from Agilent Technologies to D.S. All of the instrumentation used in this work was provided by Agilent Technologies, and the LC Image software used for presenting 2D chromatograms was provided by GC Image. Davy Guillarme wishes to thank the Swiss National Science Foundation for financial support (31003A159494).

REFERENCES

(15) Williams, A.; Read, E. K.; Agarabi, C. D.; Lute, S.; Brorson, K. A. Journal of Chromatography B. 2017, 1046, 122–130. (16) Sandra, K.; Steenbeke, M.; Vandenheede, I.; Vanhoenacker, G.; Sandra, P. Journal of Chromatography A. 2017, 1523, 283– 292. (17) Sandra, K.; Vanhoenacker, G.; Vandenheede, I.; Steenbeke, M.; Joseph, M.; Sandra, P. Journal of Chromatography B. 2016, 1032, 119–130. (18) Sorensen, M.; Harmes, D. C.; Stoll, D. R.; Staples, G. O.; Fekete, S.; Guillarme, D.; Beck, A. mAbs. 2016, 8, 1224– 1234. (19) Stoll, D. R.; Groskreutz, S. R. In Hydrophilic interaction chromatography: a guide for practitioners; Wiley, 2013; pp. 265–305. (20) Kalili, K. M.; de Villiers, A. Journal of Chromatography A. 2013, 1289, 69–79.

(1)

Strohl, W. R. Protein & Cell. 2018, 9, 86–120.

(2)

Kaplon, H.; Reichert, J. M. mAbs. 2018, In Press.

(3)

Fekete, S.; Guillarme, D.; Sandra, P.; Sandra, K. Analytical Chemistry. 2016, 88, 480–507.

(21) Pirok, B. W. J.; Gargano, A. F. G.; Schoenmakers, P. J. Journal of Separation Science. 2017.

(4)

Beck, A.; Wurch, T.; Bailly, C.; Corvaia, N. Nature Reviews Immunology. 2010, 10, 345–352.

(5)

Mullard, A. Nature Reviews Drug Discovery. 2018, 17, 86–86.

(6)

Stoll, D. R.; Zhang, K.; Staples, G. O.; Beck, A. In Advances in Chromatography; 2018.

(22) Stoll, D. R.; Sajulga, R. W.; Voigt, B. N.; Larson, E. J.; Jeong, L. N.; Rutan, S. C. Journal of Chromatography A. 2017, 1523, 162–172. (23) Stoll, D. R.; Shoykhet, K.; Petersson, P.; Buckenmaier, S. Analytical Chemistry. 2017, 89, 9260–9267. (24) Vanhoenacker, G.; Vandenheede, I.; David, F.; Sandra, P.; Sandra, K. Analytical and Bioanalytical Chemistry. 2015, 407, 355–366.

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

(25) Kalili, K. M.; de Villiers, A. Journal of Chromatography A.(29) 2013, 1289, 58–68. (26) Vivó-Truyols, G.; van der Wal, S.; Schoenmakers, P. J. Anal. Chem. 2010, 82, 8525–8536. (27) Oda, Y.; Asakawa, N.; Kajima, T.; Yoshida, Y.; Sato, T. Journal of Chromatography A. 1991, 541, 411–418. (28)

Stoll, D. In Handbook of Advanced Chromatography /Mass Spectrometry Techniques; Holcapek, M.; Byrdwell, W. C., Eds.; Elsevier: London, 2017; pp. 227–286.

Sjögren, J.; Olsson, F.; Beck, A. The Analyst. 2016, 141, 3114– 3125.

(30)

Ruta, J.; Rudaz, S.; McCalley, D. V.; Veuthey, J.-L.; Guillarme, D. Journal of Chromatography A. 2010, 1217, 8230– 8240.

(31)

Liu, B.; Guo, H.; Xu, J.; Qin, T.; Xu, L.; Zhang, J.; Guo, Q.; Zhang, D.; Qian, W.; Li, B.; Dai, J.; Hou, S.; Guo, Y.; Wang, H. mAbs. 2016, 8, 1107–1117.

(32)

Choikhet, K.; Glatz, B.; Rozing, G. LCGC Europe. 2003, 2–9.

TOC Graphic

ACS Paragon Plus Environment

Page 8 of 8