Application of a Quantitative LC–MS Multiattribute ... - ACS Publications

Feb 15, 2017 - ACS eBooks; C&EN Global Enterprise .... In this paper we describe the application of an MAM based method for site specific ... Monitori...
0 downloads 0 Views 910KB Size
Subscriber access provided by Iowa State University | Library

Article

Application of a quantitative LC-MS multi-attribute method for monitoring site-specific glycan heterogeneity on a monoclonal antibody containing two N-linked glycosylation sites Tian Wang, Lily Chu, Wenzhou Li, Kenneth A. Lawson, Izydor Apostol, and Tamer Eris Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b04856 • Publication Date (Web): 15 Feb 2017 Downloaded from http://pubs.acs.org on February 17, 2017

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 7

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 of a quantitative LC-MS multi-attribute method for monitoring site-specific glycan heterogeneity on a monoclonal antibody containing two N-linked glycosylation sites Tian Wang*, Lily Chu, Wenzhou Li, Ken Lawson, Izydor Apostol, and Tamer Eris Process Development, Amgen Inc., Thousand Oaks, CA 91320, USA ABSTRACT: A significant challenge of traditional glycan mapping techniques is that they do not provide site-specific glycosylation information. Therefore, for proteins containing multiple glycosylation sites, the individual glycan species present at a particular site cannot be differentiated from those species present at the other glycosylation sites on the molecule. Quantification of glycoform has previously been demonstrated using a multi-attribute method (MAM), which can quantify multiple post-translational modifications including deamidation, oxidation, glycosylation variants, and fragmentation1. In this paper we describe the application of an MAM based method for site specific quantification of N-linked glycan heterogeneity present on an IgG1 mAb molecule containing two distinct N-linked glycosylation sites; one present on the heavy chain (HC) variable region (Fab) and the other present on the conserved HC constant region (Fc). MAM is a peptide mapping method utilizing mass spectrometry to detect and quantify specific peptides of interest. The ionization properties of the glycopeptides with different classes of glycan structural variants, including high mannose, sialylated, and terminal galactosylated species were studied in detail. Our results demonstrate that MAM quantification of individual glycan species from both the Fab and Fc N-Linked glycosylation sites is consistent with quantification using the traditional hydrophilic interaction liquid chromatography (HILIC) analysis of enzymatically released and fluorescently labelled glycans. Furthermore, no significant impact from the glycoform on the ionization properties of the glycopeptide is observed. Our work demonstrates that the MAM method is a suitable approach for providing quantitative, site-specific glycan information for profiling of N-linked glycans on immunoglobulins.

chromatography 16,17. It is common to have co-elution of more than a single glycan species under a single chromatographically resolved peak. Chromatographically resolved glycans are identified by mass spectrometry 18,19. However, during routine analysis, peak assignment is based solely on comparison of peak retention times in the sample against peaks in a previously characterized reference standard. This approach is well accepted by regulatory agencies for routine lot release testing, but can be challenging and prone to error during early phase development where screening of a wide array of cell culture conditions and expressions systems can generate samples with diverse and variable glycan distribution. In addition, for glycoproteins with more than a single glycosylation site, the information regarding the site-specific glycan heterogeneity is lost after the glycans are released from the glycoprotein. The glycan profile obtained from the tradititional approach is from the pool of oligosaccharides released from all sites. The protein used in this study is an IgG1 mAb expressed in a SP2/0 mammalian (murine myeloma) cell line with two occupied N-linked glycosylation sites on the heavy chain (HC); one within the variable (Fab) region and the other within the conserved Fc CH2 domain. The oligosaccharides located at the two sites are a heterogeneous mixture of different glycan structures, making it challenging for site specific profiling with a conventional glycan method. Therefore, a quantitative LC-MS peptide mapping based multi-attribute method (MAM) was developed to specifically quantify glycopeptides and provide site specific

R

ecombinant monoclonal antibodies (mAbs) are increasingly prevalent therapeutic proteins, with IgG1 and IgG2 isotypes being the most common 2-5. Glycosylation is one of the most common protein post-translational modifications, with native IgG molecules containing a conserved glycosylation site in the Fc domain, and potentially additional glycosylation sites in the variable regions. IgG glycans are heterogeneous, and glycoforms vary widely between different cell lines, clones, and manufacturing conditions 6-9 (also referred to as microheterogeneity). These variations often affect biological activity, stability, immunogenicity, and half-life 10-14. As several glycan species are considered critical attributes for many therapeutic antibodies, profiling glycans in therapeutic glycoproteins is essential for controlling product consistency. Monitoring glycan profiles are also an important factor for biosimilar product development, where biologically relevant attributes such as glycans are assessed to demonstrate analytical and functional similarity to the innovator product. The traditional quantification strategy for analyzing glycans is to release the oligosaccharides from the glycoprotein enzymatically, followed by labeling of the released glycans with a fluorophore to increase detection sensitivity and/or to modify their physicochemical properties, and finally generating a chromatographic profile for identification and quantification of the released oligosaccharides. The fluorescently labeled glycans are typically quantified by fluorescence detection after separation by reversed phase (RP) 15 or normal phase liquid

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

Page 2 of 7

centration > 3.9 mg/mL) were mixed with 3.0 µL 1 M Tris (pH 7.5), 1.5 µL PNGase F (15 mU) to a total volume 30 µL. The mixture was incubated at 37 °C for 1 h. Released glycans were labeled with 50 µL of 2AA reagent (12 mg/mL 2AA, 0.2 M NaCNBH3, 0.8% (w/v) sodium acetate, and 0.4% (w/v) boric acid solution in methanol) at 80 °C for 1 h. Labeled glycans were separated from precipitated proteins by centrifugation at 12,000 rpm for 5 min. The supernatants were transferred to low volume HPLC vials for HILIC analysis. Fab Domain Purification. Reconstitute 5000 units of fabRICATOR® with 75 µL H2O (67 units/µL), add 7.5 µL (500 units) of fabRICATOR® to 500 µg of the mAb (concentration > 10 mg/mL) in cleavage buffer (10 mM sodium phosphate, pH 7.4), incubate at 37 °C for 30 min. The released Fab and Fc domain were separated by a protein A column (POROS A 20, 4.6 mm x 50 mm) on Agilent 1200 HPLC system. Flow rate was 3.0 mL/min, gradient 100% A (20 mM Tris, 150 mM NaCl, pH 7.0) from 0 – 1.4 min, 100% B (0.1 M acetic acid, 150 mM NaCl, pH 2.5) from 1.5 – 3.4 min, and 100% A from 3.5 – 5.0 min, and injection amount ≤ 200 μg protein. Fab was eluted at 0.2 – 0.4 min (Fc at approximately 2.0 – 2.2 min). Amicon Ultra-4 filter (4 mL, 10,000 MWCO, EMO Millipore, UFC801024) were used to centrifuge the Fab elution at 4000 rpm for 30 min. Fab Glycan Release from Fab Domain of the mAb. N-linked glycans on the Fab domain of the mAb were released by denaturation with 1.0 % (w/v) SDS and 50 mM DTT prior to PNGase F digestion. A total of 100 µg of the purified Fab domain of the mAb samples (concentration > 9.5 mg/mL) were mixed with 3.0 µL 5.0% (w/v) SDS and 1.5 µL 0.5 M DTT, and brought to 15 µL by adding water. The mixture was incubated at 100 °C (boiling water) for 10 min. After letting the samples cool to room temperature, 2.0 µL 15% (w/v) NP-40 solution, 3.0 µL 1 M Tris-HCl pH 7.5, 7.0 µL purified water and 1.5 µL PNGase F were added, vortexed briefly and then spun down in mini centrifuge before incubating at 37 °C for 2 h. The released glycans were labeled with 50 µL of 2AA reagent (12 mg/mL 2AA, 0.2 M NaCNBH3, 0.8% (w/v) sodium acetate, and 0.4% (w/v) boric acid solution in methanol) at 80 °C for 1 h. Labeled glycans were separated from precipitated proteins by centrifugation at 12,000 rpm for 5 min. The supernatants were transferred to low volume HPLC vials for HILIC analysis. HILIC Glycan Map. HILIC was performed using a Waters UPLC system equipped with a fluorescence detector. The 2AA- labeled N-glycans were separated on a Waters BEH Glycan column (2.1 X 100 mm, 1.7 µm) with a column temperature of 45 °C with a flow rate of 0.5 mL/min except the step specified below. Mobile phase A 50 mM ammonium formate pH 3.0 and mobile phase B 100% ACN were used. The 2AAlabeled N-glycans (2.5 µL, 3.1 µg) were injected and eluted using a gradient of 78% B from 0 – 5.0 min; 78% – 58% B from 5.0 – 41.0 min; 0% B from 41.1 – 42.9 min at a flow rate 0.4 mL/min; 78% B from 43.5 – 49.0 min. Fluorescence detection was performed with excitation at 360 nm and emission at 425 mm. Chromatographically resolved glycans are identified by mass spectrometry.

information about glycan distribution. Although quantification of glycoforms has previously been demonstrated using various LC-MS peptide mapping approaches 1,20-22 in this paper we comprehensively studied glycopeptide ionization and quantification of common Fc domain neutral glycans and Fab domain charged sialylated glycans. During cell line development for an IgG1 monoclonal antibody, clones with a wide range of Fab neutral and sialylated glycan species became available which created a unique opportunity to study the impact of glycosylation on glycopeptide quantification using MAM. We investigated in great detail the impact of the neutral and sialylated glycan composition on ionizations properties of glycopeptides which ultimately may influence the accuracy of the quantification. Especially the negatively charged, sialylated oligosaccharides may have a potential to influence charge distribution. The MAM quantification was extensively compared to the conventional hydrophilic interaction liquid chromatography (HILIC) method to assess the accuracy of the MAM results. LC-MS MAM is a derivative of conventional LC-MS/MS peptide mapping method. LC-MS/MS peptide mapping analysis of enzymatically digested proteins is routinely used in the biopharmaceutical industry for primary sequence characterization, post-translational modification quantification, and glycan profiling of biotherapeutic protein products 22,23. LC-MS MAM uses only MS1 for Pinpoint quantification which simplifies the data analysis workflow. Accurate peptide identification and quantification is enabled by application of a high resolution orbitrap mass spectrometer with accurate mass (< 5 ppm).

 EXPERIMENTAL PROCEDURES Materials. The protein used in this study is a recombinant chimeric human/mouse IgG1 mAb, which was expressed in a murine cell line SP2/0 at Amgen Inc. (Thousand Oaks, CA), and purified by standard manufacturing procedures 24. Tris(hydroxymethyl)aminomethane hydrochloride (Tris-HCl) solution, 1.0 M, pH 7.5, and pH 8.0, were produced by Teknova (Hollister, CA). fabRICATOR® (IdeS), 2-aminobenzoic acid (2AA), ammonium hydroxide (NH4OH), methanol (CH3OH), and sodium cyanoborohydride (NaCNBH3) were purchased from Sigma Aldrich (St. Louis, MO). Formic acid (FA), dithiothretol (DTT), and POROS A/20 Protein A column (4.6 mm x 50 mm) were obtained from Thermo Fisher Scientific (Waltham, MA). Iodoacetamide (IAM) and guanidine hydrochloride (GuHCl) were from MP Biomedical (Irvin, CA). NGlycosidase F (PNGase F) was purchased from Prozyme (San Leandro, CA, USA). Trypsin was obtained from Roche Diagnostic (Indianapolis, IN, USA). Acetonitrile (ACN) was a J. T. Baker reagent (Phillipsburg, NJ). ACQUITY UPLC BEH Glycan column (2.1 x 100 mm, 1.7 µm particle) and BEH Phenyl column (2.1 x 150 mm, 1.7 µm particles) were purchased from Waters Corporation (Milford, MA). Fc Glycan Release from the mAb. N-linked glycans on the mAb Fc domain from the mAb were released enzymatically by PNGase F under non-denaturing conditions (Note: Fab glycans are not released under the non-denaturing conditions; releasing of Fab glycans require sample denaturation prior to PNGase F digestion). A total of 100 µg of mAb samples (con-

2

ACS Paragon Plus Environment

Page 3 of 7

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

LC-MS Multi-Attribute Method. A total of 15 µg (concentration > 1.0 mg/mL) of the mAb was denatured and reduced by mixing with 15 µL of a denaturing solution (7.5 M GuHCl, 0.1 M Tris-HCl, pH 8.3) and 0.5 µL 0.5 M DTT. The total volume was brought to 31 µL with water. The mixture was incubated at 60 °C for 30 min. Each sample was then alkylated by adding 1.0 µL 0.5 M IAM and incubated at 25 °C for 30 min in the dark. The denatured and alkylated samples were then diluted with 62 µL 50 mM Tris-HCl pH 7.5, and trypsin digested by adding 2.5 µL trypsin (1.0 µg/µL), 0.5 µL 0.5 M DTT, and incubated at 37 °C for approximately 5 h or overnight. The digestion was quenched by lowering pH to approximately pH 2 with 1.0 µL 98% (w/v) formic acid. The samples were stored at 4 °C for subsequently LC/MS analysis. LC/MS was performed on a Waters UPLC system equipped with a UV detector and directly connected to a Thermo Scientific Q Exactive mass spectrometer equipped with an electrospray ionization (ESI) source. The high resolution, high mass accuracy and high scan speed of this instrument enable confident identification and quantification of glycopeptides Tryptic peptides of the digested antibody (7 µg, 45 µL injection) were separated on a Waters BEH Phenyl column (2.1 X 150 mm, 1.7 µm) at 50 °C with a flow rate of 0.2 mL/min. Mobile phase A 0.1% (v/v) of formic acid in water and mobile phase B 0.1% (v/v) of formic acid in ACN were used for chromatographic separation. Peptides were eluted with a gradient of 0.5% B from 0 – 5 min; 2% – 18% B from 5 – 40 min; 18% – 99% B from 40 – 45 min; 0.5% B column re-equilibration for 15 min. MS data was collected in the positive ion mode. The capillary temperature was set to 240 °C to minimize in source glycan fragmentations. The mass spectrometer was set up to collect only full MS scan with a resolving power of 70,000. The MS data from the peptide maps were analyzed in an automated fashion using Pinpoint software (Thermo Fisher Scientific, Inc). Specifically, a list of glycopeptides was selected from a reference standard injection in each sequence run using the screening function in Pinpoint. Pinpoint then automatically identified and integrated these glycopeptides in samples. These glycopeptides in samples were identified only when they were within a limited retention time window of the same glycopeptide in the reference standard, which minimized the chance of false identifications. Pinpoint generated a spreadsheet containing peak areas of all the selected glycopeptides in the samples. Percent of each specific glycopeptide was then calculated using peak area of a specific glycopeptide divided by the total peak area of all the glycopeptides on the same peptide, site-specific glycan results.

Fab) was approximately 50 with lower abundant glycan species observed at levels of 0.001%. Since the method was intended for quantification of glycopeptides solely, the peptide map gradient and data collection was stopped after the elution of the Fab glycopeptide region. The trypsin digestion conditions were also optimized for the detection of low abundant glycopeptides and to allow for analysis of lower concentration analytes. Trypsin digestion times of 1, 2, 3, 5 and 18 h (overnight) at 37 °C were evaluated. It was observed that the total area of recovered peptides reached a plateau after 5 h. Therefore, a trypsin digestion time of 5 h or overnight was adopted.

Figure 1. Chromatogram of MAM. Fc and Fab glycopeptide clusters are shown in the dotted boxes.

Evaluation of Impact of Glycan Composition on Accuracy of the MAM. Accuracy of the MAM was assessed by comparing MAM quantification results with conventional HILIC glycan map quantification results. Four samples (different clones) with a wide range of Fab neutral and sialylated glycan species were selected for the comparison of Fab glycans (Table 1). Glycan nomenclatures used in this paper are given in Chart 1. Using the traditional HILIC approach, the range of neutral glycan species (A2Ga2F) in the four samples was between 0.3% and 45.0%, and range of sialylated glycans (species containing “Sg”) ranged from 0.2% to 56.8%. In order to obtain site-specific Fab glycan results using the HILIC method, the four samples first had to be enzymatically digested with the fabRICATOR enzyme to cleave the Fab portion and then purified using affinity protein A chromatography (see Materials and Methods section). The purified Fab domain was then enzymatically digested with PNGase F to release the N-glycans, which were analyzed by HILIC glycan map to obtain Fab site-specific glycan quantification to compare with the MAM results. The first species that was analyzed consisted of a comparison of A2Ga2F quantification, a double α-Gal species, is shown in Figure 2a (appears as 3 data points, as 2 data points at 0.3% α-Gal are overlapping). A linear regression analysis was performed: the slope was 0.939, intercept was zero, and correlation coefficient (R2) was 0.999. The second species on the Fab site that was compared was a biantennary glycan with one terminal α-Gal and one terminal sialic acid (A2Sg1Ga1F). The

 RESULTS AND DISCUSSION Optimization of the MAM. The MAM method for quantification of Fc and Fab glycopeptides was developed based on a conventional LC-MS/MS peptide map, and used a shallow gradient to optimize the resolution and sensitivity of the glycopeptides. The Fc glycopeptides eluted within a 6 min window at approximately 20 min, while the Fab glycopeptides eluted within an 8 min window at approximately 35 min (Figure 1). The number of glycopeptides observed on each site (Fc and

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

Page 4 of 7

linear regression analysis showed a slope (Figure 2b) of 0.895, intercept of 0.014, and an R2 of 0.997. For sialylated species, in addition to the data in Figure 2b, single, double and triple sialylated species were evaluated separately (Figure 2c, 2d, and 2e). The linear regression analysis showed R2 values for all glycan species were from 0.997 to 0.999, which demonstrates good overall linearity for all glycan species. In addition, the slopes were close to one (between 0.896 and 0.947) and the intercepts were near zero for all of the glycan species, which indicates a good correlation and accuracy against the conventional HILIC results. Table 1. Glycan data of the four different samples of the Mab

Figure 2. Plots of MAM data vs. HILIC glycan map data for Fab site glycans. Panel a, biantennary α-Gal species; panel b, mono antennary α-Gal species; panel c, mono antennary sialylated species; panel d, biantennary sialylated species; and panel e, triantennary sialylated species.

Chart 1. Glycan nomenclature and structures

Fc glycopeptides also showed a similar accuracy and linearity (Figure 3). Due to specific biological relevance, subgroups of glycans, high mannose (sum of mannose 5 to mannose 8), aFucose (sum of glycans without fucose, excluding high mannose), and β-galactose (sum of glycans with terminal galactose), were analyzed and are presented in Figure 3. A large number of data points (more than 30) was included in Figure 3 since the procedure for the Fc HILIC glycan map is relatively less complex than the Fab HILIC glycan map. The conventional HILIC glycan mapping method releases the Fc glycans readily under non-denaturing conditions without releasing the Fab glycans, therefore, no Fc domain purification was required.

Figure 3. Plots of MAM data vs. HILIC glycan map data for Fc site glycans. Panel a, high mannose; panel b, aFucose; panel c, βGalactose.

It was anticipated that the degree of sialylation of a glycopeptide may influence the MS ionization properties of the glycopeptide, and therefore impact the accuracy of the MAM method for quantification of sialylated glycopeptides. To understand the accuracy of MAM, charge profiles of glycopeptides containing 0 to 3 sialic acids were further examined (Figure 4). All of glycopeptide species shown in Figure 4 had two charged forms, Z=3 and Z=4, in the m/z window of 350-2000;

4

ACS Paragon Plus Environment

Page 5 of 7

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

however the relative intensity of the 2 charged forms varied. A3Sg3F, containing 3 sialic acids had the highest relative abundance of Z=4 vs. Z=3 species, while A2Ga2F with 0 sialic acids had approximately equal abundance of Z=3 and Z=4 species. The distribution of charge species were quantified using extracted ion chromatograms of the Z=3 and Z=4 ions, and the average charge was calculated per to the equation below. Average Charge =

∑ (Peak Area X Charge ) Total Peak Area

Figure 5. Plot of average charges of glycopeptides vs. mass of glycopeptides (a) or number of sialic acid (Sg) in glycopeptides (b). Open circles representing glycopeptides with zero Sg; open squares representing glycopeptides with 1, 2, or 3 Sg. Error bars (among the 4 different samples) are also shown in Figure 5a.

Figure 4. Mass spectra of Fab glycopeptides. The charged glycopeptides (in the boxes) with number of charges shown at the boxes.

The plot of average charge vs. peptide mass (Figure 5a) from the four samples tested (Table 1) showed a strong linear relationship (R2 of 0.96). The aglycopeptide (lacking glycosylation) and smaller asialoglycopeptides (A1G0F, A2G0F, A2G1F, and A2G2F) were also included in the evaluation to better understand the impact of peptide mass to the average charge. An average charge of 2.67 (from the four different samples of the mAb) was observed for the aglycopeptide with a mass of 1906 Da and an average charge of 3.87 for the largest glycopeptide, A3Sg3G3F, with a mass of 4962 Da. Glycopeptides without any sialic acid (SA 0, open circles in Figure 5a) had increasing average charge with increasing mass, from 2.67 average charge for the aglycopeptide to 3.50 average charge for the A2Ga2F glycopeptide. The larger sialylated glycopeptides, containing 1, 2, or 3 sialic acids (Sg 1, 2, 3; open squares in Figure 5a), also showed increasing average charge with increasing mass. All glycopeptides, Sg 0, 1, 2, or 3, fit with high correlation into the same trend line (Figure 5a). This indicates that the average charge of glycopeptides is not a function of number of sialic acids on the glycopeptide, but a function of the mass of the glycopeptides. Data in Figure 5a was replotted using average charge vs. number of Sg (Figure 5b). It shows again that the average charge of glycopeptides is not a function of the number of sialic acids.

A plot of average charge vs. relative abundance of the glycopeptides in the four samples of the mAb (Table 1) showed flat relationship (Figure 6). As shown in Figure 6, the relative abundance of the A2Sg2F glycopeptide in the samples varied from 3% to 46%, with the average charge values of 3.72 to 3.76 across the entire range. Similar results were obtained for the other glycopeptides plotted in Figure 6. These flat lines may play a role in the excellent linearity previously observed for glycopeptide quantification using MAM (Figure 2 and Fig 3). Precision of the method was assessed from 11 independent injections of the mAb reference standard. The RSD for Fc glycans: %high mannose, %aFucose, and %β-Galactose were 5%, 11%, and 3%, respectively. These results are equivalent to those obtained with the Fc-HILIC glycan mapping method of 7%, 11%, and 3%, respectively. The RSD for Fab glycans: %α-Galactose and total %sialylated species were 2% and 5%, respectively, which are similar to that of the major Fc glycopeptides (high mannose and β-Galactose). The RSD from the FabHILIC glycan mapping method was not available for direct comparison with the MAM due to complexity of the FabHILIC glycan map procedure and limited number of assays performed.

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

Page 6 of 7

information for each glycosylation site, and were superior to the conventional HILIC assay which required a laborious enzymatic digestion and chromatographic separation step prior to analysis in order to obtain the same, site specific information as MAM. Our work has demonstrated that the quantitative MAM method using automated software (PinPoint) is suitable for providing site-specific glycan information for profiling of Nglycans on glycoproteins. This high resolution analysis is an important tool for monitoring biologically relevant product attributes during process development.

 ABBREVIATIONS ACN: DTT: FA: Fab: Fc: GuHCl: HILIC:

Figure 6. Plot of average charges of glycopeptides vs. abundance of glycopeptides.

In addition to MAM, recent advances in other MS quantification techniques including multiple reaction monitoring (MRM) 20,21 and MALDI-TOF 25,26 have shown promise for the quantification of glycopeptides. MRM is a targeted method which monitors pre-defined m/z values for the glycopeptide species of interest. MRM has high sensitivity, and is useful for analysis of known or expected glycopeptides in complex mixtures such as serum. However MRM cannot detect unknown or unexpected species, which can limit comprehensive, retrospective analysis of the sample for all potential glycopeptides. Retrospective analysis of the sample is possible using the nontargeted LC/MS data acquisition employed for MAM. All m/z values generated within the specified range are recorded during analysis so it is possible to search for and quantify unknowns. The sensitivity of MAM is also very high, with quantification of species at 0.1% relative abundance, mainly because the analysis is performed on a single purified protein and not in a complex mixture such as serum. MALDI-TOF is another glycopeptide quantification approach and is useful for detecting peptides that are lost in HPLC in some cases 25. However MALDI-TOF is less quantitative and less sensitive than LC-MS based methods due to lack of HPLC separation of the peptides which can lead to ion suppression 26. The results from our MAM study demonstrated good accuracy and precision compared to the conventional glycan quantification approach, whereas glycopeptide quantification by MRM and MALDI-TOF requires further evaluation.

IAM: LC: mAbs: MAM: MS: RT: Tris: 2AA:

acetonitrile dithiothreitol formic acid antibody variable domain antibody conserved CH2 and CH3 domains guanidine hydrochloride hydrophilic interaction liquid chromatography iodoacetamide liquid chromatography monoclonal antibody(ies) multi-attribute method mass spectrometry room temperature tris(hydroxymethyl)aminomethane base 2-Aminobenzoic acid

 AUTHOR INFORMATION *Corresponding Author Email: [email protected]

 ACKNOWLEDGEMENT We thank Dr. S. Benchaar for reviewing part of the data.

 REFERENCES (1) Rogers, R. S.; Nightlinger, N. S.; Livingston, B.; Campbell, P.; Bailey, R.; Balland, A. MAbs 2015, 7, 881-890. (2) Reichert, J. M. Nat Biotechnol 2001, 19, 819-822. (3) Reichert, J. M. Curr Opin Mol Ther 2002, 4, 110-118. (4) Reichert, J.; Pavolu, A. Nat Rev Drug Discov 2004, 3, 383-384. (5) Reichert, J. M.; Rosensweig, C. J.; Faden, L. B.; Dewitz, M. C. Nat Biotechnol 2005, 23, 1073-1078. (6) Patel, T. P.; Parekh, R. B.; Moellering, B. J.; Prior, C. P. The Biochemical journal 1992, 285 ( Pt 3), 839-845. (7) Jenkins, N.; Parekh, R. B.; James, D. C. Nat Biotechnol 1996, 14, 975-981. (8) Hills, A. E.; Patel, A.; Boyd, P.; James, D. C. Biotechnol Bioeng 2001, 75, 239-251. (9) Baker, K. N.; Rendall, M. H.; Hills, A. E.; Hoare, M.; Freedman, R. B.; James, D. C. Biotechnol Bioeng 2001, 73, 188-202. (10) Arnold, J. N.; Wormald, M. R.; Sim, R. B.; Rudd, P. M.; Dwek, R. A. Annual review of immunology 2007, 25, 21-50. (11) Raju, T. S. Current opinion in immunology 2008, 20, 471-478.

 CONCLUSIONS A quantitative multi-attribute method (MAM) was developed to monitor glycan profiles of two distinct glycosylation sites on a therapeutic immunoglobulin. Our results demonstrated that quantification of the glycopeptides using an MAM based approach provided equivalent results to those obtained using the conventional HILIC glycan mapping technique with similar accuracy and precision. A detailed study of quantification of heterogeneous glycopeptides indicated that negatively charged, sialylated glycans had similar ionization properties and quantification as the neutral glycopeptides, and did not impact the accuracy of glycopeptide quantification by MAM. The MAM results provided the glycan micro-heterogeneity

6

ACS Paragon Plus Environment

Page 7 of 7

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

(12) Goetze, A. M.; Liu, Y. D.; Arroll, T.; Chu, L.; Flynn, G. C. Glycobiology 2012, 22, 221-234. (13) Goetze, A. M.; Liu, Y. D.; Zhang, Z.; Shah, B.; Lee, E.; Bondarenko, P. V.; Flynn, G. C. Glycobiology 2011, 21, 949-959. (14) Lifely, M. R.; Hale, C.; Boyce, S.; Keen, M. J.; Phillips, J. Glycobiology 1995, 5, 813-822. (15) Chen, X.; Flynn, G. C. Analytical Biochemistry 2007, 370, 147161. (16) Anumula, K. R. Anal Biochem 2006, 350, 1-23. (17) Anumula, K. R. Glycobiology 2012, 22, 912-917. (18) Wuhrer, M.; de Boer, A. R.; Deelder, A. M. Mass Spectrometry Reviews 2009, 28, 192-206. (19) Wuhrer, M.; Deelder, A. M.; Hokke, C. H. Journal of Chromatography B 2005, 825, 124-133. (20) Hong, Q.; Lebrilla, C. B.; Miyamoto, S.; Ruhaak, L. R. Anal Chem 2013, 85, 8585-8593.

(21) Song, E.; Zhu, R.; Hammoud, Z. T.; Mechref, Y. Journal of proteome research 2014, 13, 4808-4820. (22) Shah, B.; Jiang, X. G.; Chen, L.; Zhang, Z. Journal of the American Society for Mass Spectrometry 2014, 25, 999-1011. (23) Traylor, M. J.; Tchoudakova, A. V.; Lundquist, A. M.; Gill, J. E.; Boldog, F. L.; Tangarone, B. S. Anal Chem 2016, 88, 9309-9317. (24) Shukla, A. A.; Hubbard, B.; Tressel, T.; Guhan, S.; Low, D. J Chromatogr B Analyt Technol Biomed Life Sci 2007, 848, 28-39. (25) Plomp, R.; Hensbergen, P. J.; Rombouts, Y.; Zauner, G.; Dragan, I.; Koeleman, C. A.; Deelder, A. M.; Wuhrer, M. Journal of proteome research 2014, 13, 536-546. (26) Hale, J. E.; Butler, J. P.; Knierman, M. D.; Becker, G. W. Anal Biochem 2000, 287, 110-117.

Table of Contents artwork

7

ACS Paragon Plus Environment