Subscriber access provided by University of Newcastle, Australia
Article
Hydrophilic Interaction Chromatography (HILIC) hyphenated with Mass Spectrometry: a powerful analytical tool for the comparison of originator and biosimilar therapeutic monoclonal antibodies at the middle-up level of analysis. Valentina D'Atri, Szabolcs Fekete, Alain Beck, Matthew Allen Lauber, and Davy Guillarme Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b04726 • Publication Date (Web): 12 Jan 2017 Downloaded from http://pubs.acs.org on January 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 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
Hydrophilic Interaction Chromatography (HILIC) hyphenated with Mass Spectrometry: a powerful analytical tool for the comparison of originator and biosimilar therapeutic monoclonal antibodies at the middle-up level of analysis. Valentina D’Atri,*,† Szabolcs Fekete,† Alain Beck,‡ Matthew Lauber,§ and Davy Guillarme.† † ‡ §
School of Pharmaceutical Sciences, University of Geneva, University of Lausanne, Geneva, Switzerland. Center of Immunology Pierre Fabre, 5 Avenue Napoléon III, BP 60497, Saint-Julien-en-Genevois, France. Waters Corporation, 34 Maple Street, Milford, Massachusetts 01757-3696, United States.
ABSTRACT: The development and approval processes of biosimilar mAbs depend on their comparability to originators. Therefore, analytical comparisons are required to assess structural features and post-translational modifications (PTM) and thereby minimize the risk of being clinically meaningful differences between biosimilar and originator drug products. The glycosylation pattern of mAbs is considered to be an important critical quality attribute (CQA), and several analytical approaches have been proposed that facilitate characterizing and monitoring a glycosylation profile, albeit mainly at a glycan and glycopeptide level of analysis. In this study, we demonstrate the utility of hydrophilic interaction chromatography (HILIC) hyphenated with mass spectrometry (MS) for the qualitative profiling of glycosylation patterns at the protein level, by comparing originator and biosimilars mAbs (Remicade®/Remsina®/Inflectra®, Herceptin®/Trastuzumab B, and Erbitux®/Cetuximab B) using a middle-up approach. We demonstrate the ability of HILIC to resolve hydrophilic variants of protein biopharmaceuticals at the middle-up level of analysis, its complementarity to reversed phase liquid chromatography (RPLC), and its hyphenation to MS. HILIC features combined to MS make a powerful analytical tool for the comparison of originator and biosimilar mAbs that could eventually be applied in routine analyses for quality control.
Monoclonal Antibodies (mAbs) represent the largest and fastest growing category of biopharmaceutical proteins.1 They are manufactured through recombinant DNA technology, custom-designed to target specific antigens and employed for a wide range of applications, particularly in cancer, immune disorders, and infectious diseases.2,3 Furthermore, mAbs constitute the largest market segment of current biopharmaceutical products. In addition, because several commercial mAbs are approaching patent expiry,4,5 a market for so-called biosimilars is becoming increasingly attractive and is accordingly undergoing rapid growth.6 Before being approved for clinical use, it must be demonstrated that a biosimilar is highly similar to the reference originator, though it may be that there are some minor differences that need to be carefully tracked and tested for any correlation to bioactivity and safety. In sum, there is a motivation in biosimilar development to pursue comprehensive comparability studies to reduce the burden of new clinical trials. As with other biopharmaceuticals, mAbs and their related biosimilars exhibit structural micro-heterogeneity, mainly due to their expression from living organisms and that they are substrates for a wide range of post translational modifications (PTM). Resulting micro-variants may have potential impacts on biopharmaceutical function, stability, and efficacy, and in some cases can seriously affect patient safety.7,8 For these reasons, sophisticated analytical characterization is required to
ensure extensive and satisfactory quality control during each stage of manufacturing and process development.9 Indeed, therapeutic mAbs have a complex, heterogeneous IgG-based glycoprotein structure, containing glycans at a weight percent of about five percent (Figure 1). Despite their low percentage, glycans can impact several functions of a mAb.10 Moreover, the glycan profile of a mAb is affected by the cell line used for expression, with Chinese hamster ovary cells (CHO) and mouse melanoma-derived cells (SN2/0 or NS0) being the most widely used for manufacturing. Generally, major forms of glycans consist of complex, biantennary oligosaccharides containing N-acetylglucosamines and mannoses as components of the main core (which itself is frequently fucosylated), and zero, one, or two terminal galactose residues along with very low levels of sialylation. For an IgG manufactured in mouse-derived cell lines, N-glycolyl neuraminic acid (NGNA), the hydroxylated form of the N-acetyl neuraminic acid (NANA) expressed in humans, is preferentially expressed, together with a fraction of bisecting glycans, not present with CHO cell production.11 Several examples of mAb glycosylation patterns affecting function, stability, efficacy, and safety have been reported.7,8 For instance, afucosylated12 or galactosylated variants13–15 can lead to increased effector functions,16 high-mannose variants can accelerate clearance,17,18 and NANA sialylation can enhance antiinflammatory properties.19–21 Most importantly, NGNA sialylation and galactose-α1,3-galactose motifs might affect
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
product safety as a result of immunogenicity and causing adverse patient side effects.22,23 Due to its importance, the glycosylation pattern of a mAb is often a critical quality attribute (CQA)8 and it must be carefully monitored during all stages of manufacturing, especially during process development.24,25 Several analytical methods can be applied for the structural characterization of mAbs,26–29 and among them, HPLC is one of the most widely used. Despite the fact that complementary information on biopharmaceutical properties (such as charge, size or hydrophobic variants) can be obtained with a wide range of HPLC modes, none of these modes offers a suitable elucidation of the glycosylation profile at the protein level. In addition, besides reversed phase liquid chromatography (RPLC), other chromatographic modes are not easily compatible on-line with mass spectrometry (MS). So, despite the fact that they provide an optimal degree of separation, their suitability for high confidence peak identification remains limited.
Figure 1. a) Structural features of an IgG1-based mAb. Two identical heavy chains (in light blue) and light chains (in light grey) are held together by interchain disulfide bonds (in red). The fragment antibody region (Fab) is responsible for antigen binding through the variable regions of the heavy and light chains (vHC and vLC, respectively). The crystallisable fragment (Fc) is responsible for the interaction with cell surface receptors. HC stands for heavy chain, LC for light chain, v for variable region, c for conserved region, and Hi for hinge region. Glycosylation at a conserved site in the cHC2 domain is represented in green. In some cases, the vHC domain can also be N-glycosylated. b) Schematic representation of some of the common N-glycoforms that might be found in IgG-based mAbs. The following glycan nomenclature was used: H = Hexose (Mannose/Galactose); N = N-Acetyl Glucosamine; F = Fucose; Sa = N-Acetyl Neuraminic
Page 2 of 8
Acid (NANA); Sg = N-Glycolyl Neuraminic Acid (NGNA). Numbers indicate the composition of the indicated residue.
In this scenario, hydrophilic interaction chromatography (HILIC) presents itself as an attractive alternative. By using a polar stationary phase in combination with a highly organic mobile phase, HILIC is fully compatible with electrospray ionization mass spectrometry (ESI-MS) and offers a complementary retention mechanism compared to RPLC, since analytes are eluted based on their hydrophilicity. This chromatographic method has already been applied in the field of biopharmaceuticals for released glycan profiling and glycopeptide separations. However, its effective application to large biomolecules has only recently become possible, as we have demonstrated in a recent proof of concept paper.30 In fact, thanks to the availability of purposefully designed wide-pore HILIC phases, there are now new possibilities for the analytical characterization of glycoproteins at intact and middle-up levels of analysis (Scheme S1). In this study, we compare the capabilities of reversed-phase liquid chromatography (RPLC) and hydrophilic interaction chromatography (HILIC) to characterize biosimilar candidates relative to their respective originators, with a focus on profiling glycan compositions at the middle-up level of analysis. EXPERIMENTAL SECTION Reagents and materials. Water and acetonitrile (ACN) were UHPLC-MS grade and purchased from Biosolve (Valkenswaald, Netherlands). Trifluoracetic acid (TFA, >99%), tris(hydroxymethyl)aminomethane (TRIS), dithiothreitol (DTT), guanidine hydrochloride (GuHCl), and 1 M hydrochloric acid (HCl) solution were obtained from SigmaAldrich (Buchs, Switzerland). Immunoglobulin-degrading enzyme of Streptococcus pyogenes (IdeS, FabRICATOR®) was purchased from Genovis Inc. (Lund, Sweden). mAbs and biosimilars were kindly provided by Pierre Fabre laboratories (Saint-Julien en Genevois, France). Sample preparation. Each mAb was diluted from the stock solution to a concentration of 1 mg/mL (100µg) with UHPLCMS grade water and incubated with 100 units of IdeS enzyme in 10 mM TRIS buffer pH 7.5 at 45 °C for 45 minutes. The resulting Ides-digested mAb was then denatured and reduced by the addition of 1 M DTT and solid guanidine hydrochloride (GuHCl) for a final buffer composition of approximately 100 mM DTT, 4 M GuHCl, and 10 mM TRIS. The incubation of the IdeS-digested mAb in this buffer was performed at 45 °C for 30 minutes. The same protocol was adopted for the digestion and reduction of all biosimilars. Instrumentation and columns. LC-MS analyses were performed using an ultra-high-performance liquid chromatography (UHPLC) system (Waters ACQUITY UPLC, Milford, MA, USA), equipped with a binary pumping system and fixed loop injector. This UHPLC was coupled to an electrospray time-of-flight mass spectrometer (XevoTM Q-ToF, Waters, Milford, MA, USA) and a fluorescence detector (FD). The mass spectrometer was operated in positive ion mode and ions were scanned over an m/z range of 400-4000 with a 1 s scan rate. Capillary voltage was set to 3.0 kV, sample cone voltage to 30 V, source temperature to 150 °C, desolvation gas temperature to 500 °C and gas flow to 1000 L/h. The instrument was calibrated using the singly charged ions produced by a 2 µg/µL sodium iodide solution in 2-propanol/water (1:1). Data
ACS Paragon Plus Environment
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
acquisition and analysis were performed with MassLynx 4.1 (Waters). Deconvolutions were performed by manual identification of the multiply-charged series. Then electrospray spectra were transformed onto a molecular mass axis by specifying the resolution between data points in the transformed spectrum as 0.1 Da and by specifying lowest/highest molecular mass that the algorithm had to consider for the peak series, generally around 23000-24000 Da for LC, 24000-28000 Da for Fc/2 (depending on the glycans size attached on it), and 2500026000 Da for Fd’. Columns used in the study were ACQUITY UPLC BEH C18 300 Å 1.7 µm (2.1 mm ID x 150 mm) for RPLC and ACQUITY UPLC Glycoprotein Amide 300 Å 1.7 µm (2.1 mm ID x 150 mm) for HILIC, both obtained from Waters (Milford, MA). Chromatographic conditions. Mobile phase consisted of 0.1% of TFA (v/v) in water (A) and 0.1% TFA (v/v) in ACN (B). The column temperature was 80 °C, and injection volume was 1 µL (fixed 1 µL loop size used in full loop injection mode). For RPLC, the flow rate was set to 0.4 mL/min, and the gradient conditions consisted of 30% to 38% B in 12 minutes for the analysis of Trastuzumab and Infliximab and their respective biosimilars; and 30% to 40% B in 12 minutes for the analysis of Cetuximab and its biosimilar. In all cases, the gradient was followed by a 1 min washing step at 70% B and a 5 min re-equilibration step. Weak needle wash composition was 80% water and 20% ACN. For HILIC, the flow rate was 0.45 mL/min, and the gradient was 85% to 75% B in 0.2 min, then 75% to 65% B in 10 min for Trastuzumab and its biosimilar; 85% to 73% B in 0.2 min, then 73% to 65% B in 10 min for Cetuximab and its biosimilar; or 85% to 73% B in 0.2 min, then 73% to 69% B in 10 min for Infliximab and its biosimilars. Also in HILIC, the gradient was systematically followed by a 1 min washing step at 15% B and a 5 min reequilibration step. Weak needle wash composition was 85% acetonitrile and 15% water. Chromatographic conditions were systematically optimized during preliminary studies. RESULTS AND DISCUSSION General considerations regarding TFA and injection mechanism. In this study, the use of trifluoroacetic acid (TFA) as a mobile phase modifier was a critical compromise. TFA acts as a pH adjuster and ion-pairing reagent. For ESIMS analysis, its use is often questioned, because it causes ion suppression when compared to weaker acids, like formic acid. Nevertheless, LC-MS with a low percentage TFA mobile phase is quite feasible, given volatility of the acid.31 Most importantly, because TFA acts as an ion-pairing reagent, it generates some solvating effects that crucially enhance chromatographic performance in terms of protein separation selectivity, resolution and peak shape.30 Indeed, the concerted effects of the strongly acidic conditions imparted by TFA and its effective ion-pairing help to yield a hydrophobic shield on the glycoprotein and to minimize ionic interactions with a stationary phase. This increases the selectivity of the hydrophobic species under RPLC conditions and allows for the glycosylation pattern to more exclusively dictate the retention mechanism of hydrophilic partitioning in HILIC separation. Based on the advantages resulting from the use of TFA, we considered the use of a 0.1% TFA modified mobile phase to be an acceptable compromise for LC-MS analysis, despite the decrease in MS sensitivity. There are also a few practical aspects related to HILIC analysis of large biomolecules that need to be considered. As
previously reported,30 particular attention should be paid to the sample diluent to avoid peak distortion and band broadening.32 To prevent this issue, the sample diluent should ideally be as close as possible to the initial mobile phase composition of a separation, thus a sample diluent with an organic solvent composition equal to initial mobile phase conditions should be preferred. Unfortunately, this is not possible, especially with biopharmaceutical products, being that they can have limited solubility and be susceptible to precipitation in the presence of a high percentage of organic solvent. An aqueous sample of mAb was injected in HILIC, but the strong eluotropic strength of the sample diluent was counter-balanced by a fast initial gradient ramp that integrates a high percentage of ACN (85%) at the beginning of the method.30 In addition, a small volume was injected and the needle wash solvents (such as strong/weak needle wash, purge solvent, etc) were optimized. Prior to the analysis, it was therefore mandatory to evaluate the possible effects of the injection mode on peak shapes and preferentially set the wash solvent composition in accordance with the initial mobile phase conditions. Samples preparation prior LC-MS analyses. mAbs (and their related biosimilars) have an overall molecular mass of approximately 150 kDa, of which circa five percent corresponds to glycans. Even though the analysis of the mAbs at the intact level would represent the most effective approach for a complete characterization, it is hampered by the multisite heterogeneity of mAbs and the lack of chromatographic and/or mass spectrometric resolution. Applying a middle-up approach (digestion and reduction of the intact glycoproteins) facilitates the characterization of mAb heterogeneity and makes it possible to more accurately assess glycosylation profiles, as the analysis is conducted on subunits of around 25 kDa. In this study, the immunoglobulin-degrading enzyme of Streptococcus pyogenes was used to selectively cleave the mAb under the hinge region and to release single chain Fc fragments (Fc/2). By using dithiothreitol (DTT) as a reducing agent, light chain (LC) and Fd’ domains were also obtained so as to yield a final sample containing three main fragments of almost 25 kDa each that can be more easily analyzed by LCMS(Scheme 1). After digestion and reduction of the intact mAbs and their related biosimilars, the obtained subunits were analyzed using both RPLC and HILIC wide pore (300Å) stationary phases and separations directly coupled to an ESI-Q-TOF mass spectrometer. The obtained profiles are discussed in the following sections.
Scheme 1. Middle-up sample preparation flowchart. The intact mAb was digested using IdeS to yield F(ab)’2 and Fc/2 subunits. DTT reduction was then used to cleave the disulfide bonds of the F(ab)’2 subunit to yield Fd' and LC subunits. The final sample contains three main fragments of about 25 kDa each. For color coding, please refer to Figure 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
Infliximab. Infliximab (Remicade®) is a chimeric mAb used in the treatment of autoimmune diseases where it acts against tumour necrosis factor alpha (TNF-α), a key cytokine involved in autoimmune reactions. Inflectra® and Remsina® are the biosimilar versions of Remicade®, both of them already approved for use in Europe, South Korea and the United States.33 Inflectra® and Remsina® share an identical amino acid sequence versus the originator (Remicade®), and they generally show a very high similarity in their glycan composition, with H3N4F1 and H4N4F1 representing the most abundant moieties of the glycosylation pattern.34 Figure 2 shows the total ion chromatograms (TIC) obtained from RPLC- and HILIC-MS analyses. By hyphenating RPLC with MS (Figure 2i), and thanks to the wide pore stationary phase used in the study, it was possible to produce a separation of the three main fragments that are generated upon IdeS digestion and reduction (Fc/2, LC and Fd’ fragments, in their order of elution). Moreover, the Fc/2 fragment eluted in two peaks. The second peak was associated with the hydrophobicity change of the fragment, a consequence of Cterminal lysine cleavage. Interestingly, the double peak profile of Fc/2 for Remsina® shows it to have a significantly different extent of lysine truncation as compared to both Inflectra® and Remicade®. In any case, it should be noticed that several Fc/2 variants bearing different glycans co-eluted in these RPLC peaks, and they could only be identified with MS (see Table S1 for details). Analyzing the same samples in HILIC-MS (Figure 2ii) offers complementary information. The elution order of the fragments was reversed compared to RPLC (Fc/2 fragment eluted later than the other subunits).
Page 4 of 8
Despite the fact that the Fd’ and LC fragments were not well resolved, which suggests they have very similar hydrophilicity, the Fc/2 fragment was now resolved into two main peaks corresponding to fragments bearing the most abundant glycan moieties (H3N4F1 and H4N4F1). Moreover, additional minor peaks were observable, each corresponding to less abundant moieties, such as H3N3F1, H3N4, H5N2, H5N4F1, and in some cases to the presence of some sialylated species, such as H4N4F1Sg1 and H5N4F1Sg1 (see Table S2 for details). Interestingly, the selectivity of the HILIC separation makes it possible to directly assess qualitative differences in the glycosylation of the originator and biosimilar mAbs. In this example, it is possible to discriminate the different amounts of H4N4F1 between Infliximab, Inflectra® and Remsina® (Figures 2aii, 2bii, and 2cii respectively, focus on the second major peak related to the Fc/2 fragment), which is consistent with the work by Sorensen et al.35 where a comparison of Remicade® and Remsina® was performed by a more complex 2D-LC-MS setup. It should be noted that, in the HILIC separation, the Fc/2 lysine variants co-eluted, since no separation was obtained based on electrostatic interactions by virtue of TFA ion pairing. We should point out that one of the aims of this study was to prove the complementarity of HILIC vs RPLC by also highlighting the different selectivity offered by the two methods. For this reason, samples containing Fc/2 lysine variants were purposefully analyzed. Trastuzumab. Trastuzumab (Herceptin®) is a humanized mAb used for the treatment of certain types of breast cancer. It acts by interfering with the signalling pathway responsible for cancer cell proliferation as mediated by the human epidermal growth factor receptor 2 (HER2).
Figure 3. Middle-up analysis of (a) Trastuzumab (Herceptin®), and its biosimilar: (b) Trastuzumab B. Total ion chromatograms of the (i) RPLC-MS and (ii) HILIC-MS analyses. See Tables S3 and S4 for detailed retention times and mass assignments.
®
Figure 2. Middle-up analysis of (a) Infliximab (Remicade ), and its biosimilars: (b) Inflectra® and (c) Remsina®. Total ion chromatograms of the (i) RPLC-MS and (ii) HILIC-MS analyses. See Table S1 and S2 for detailed retention times and mass assignments.
Trastuzumab B differs from its originator by only one amino acid residue in the heavy chain. This amino acid substitution occurs in the Fd’ domain and involves two charged residues (R217K), thus no major change in retention time is expected at the chromatographic level, but instead only a shift of +28 Da as detected by MS.36 Moreover, because both origina-
ACS Paragon Plus Environment
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
tor and biosimilar are expressed in a CHO cell line, the glycosylation pattern is expected to be highly similar. Figure 3 shows chromatograms obtained from RPLC- and HILIC-MS analysis of Trastuzumab (Herceptin®) and Trastuzumab B. As with Infliximab, RPLC-MS analysis (Figure 3i) afforded a separation of the three main fragments but
offered information about the glycosylation pattern only after interrogation of the collected MS data (data not shown, see Tables S3 for details). In contrast, HILIC-MS analysis performed on the same samples (Figure 3ii) allowed for a direct and immediate comparison of the glycosylation profiles.
Figure 4. Middle-up RPLC-MS analysis of (a) Cetuximab (Erbitux®) and (b) Cetuximab B. Total ion chromatograms (i, v) and associated deconvoluted mass spectra of Fc/2 (ii, vi), LC (iii, vii), and Fd’ (iv, viii) fragments. For the sake of simplicity, only the deconvoluted mass spectra of the second peak of the Erbitux® Fc/2 fragment is shown. Middle-up HILIC-MS analysis of Erbitux® (c) and Cetuximab B (d). Fc/2 fragment assignments are accompanied by “+” or a red asterisk depending on whether or not the eluting species contains its Cterminal lysine residue (respectively). See Tables S5 and S6 for details on retention times and mass assignments.
The two main peaks of the Fc/2 fragments correspond to H3N4F1 and H4N4F1 glycoforms, while the additional minor peaks are related to H3N3F1, H5N4F1, H3N4, and H4N4 glycoforms. It should point out that, in contrast with Infliximab HILIC profiles (Figure 2ii), the resolution on the H4N4F1 isomers is more evident in these chromatograms, because there is no distortion from lysine variant content. The glycosylation profile of Trastuzumab B also shows the afucosylated H5N2. No sialylated species were detectable in either sample (see Table S4 for details). Cetuximab. Cetuximab (Erbitux®) acts as epidermal growth factor receptor (EGFR) inhibitor, and it is a chimeric mAb used for the treatment of several types of metastatic cancer. This mAb represents quite an intriguing example that highlights the capability of HILIC-MS for qualitative comparability studies of originator/biosimilar glycosylation patterns. In contrast with most other therapeutic mAbs, Erbitux® exhibits a second N-glycosylation site on the variable portion of its heavy chain (vHC, Figure 1). This second glycosylation site exhibits greater glycan heterogeneity that can in turn cause for hypersensitivity reactions in patients.7,9,22,36,37 Being produced in a murine SP2/0 cell line, the glycosylation sites of Erbitux®
have a propensity for accommodating hyper-galactosylated structures and species containing N-Glycolyl Neuraminic Acid (NGNA, denoted as Sg in Figure 1b); bisecting Nacetylglucosamines might also be expected.38–40 The biosimilar version, namely Cetuximab B, shares the same amino acid sequence as the originator, but it displays a different glycosylation pattern due to its expression from a CHO cell line. The LC-MS separations of Erbitux® and Cetuximab B following IdeS enzymatic digestion and reduction with dithiothreitol (DTT) are shown in Figure 4. Total ion chromatograms (TIC) obtained from an RPLC-MS analysis (Figures 4i and 4v) highlight the separation of the three main fragments: Fc/2, LC and Fd’. As with Infliximab, the double peak of the Fc/2 fragment was found to be associated to C-terminal lysine truncation. Thanks to the deconvolution of the mass spectra corresponding to the Fc/2 (Figures 4ii, 4vi) and Fd’ (Figures 4iv, 4viii) fragments, it was possible to define the differences occurring in the glycosylation profiles and highlight the glycan heterogeneity present at the Fd’ Nglycosylation site (Figures 4iv, 4viii and see Table S5 for details).
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
A more direct visualization of this heterogeneity was offered by the HILIC-MS analysis (Figures 4c, 4d, S1, S2). In contrast to the other HILIC profiles obtained in this work, the light chains (denoted as LC) eluted first, with the glycosylated fragments eluting just after. The first retention window of glycosylated fragments was confirmed by MS to correspond to Fc/2 glycoforms and the second retention window to Fd’ glycoforms. It should be noted that, as for Infliximab (Figure 2ii), the Fc/2 lysine variants co-eluted (Figure S1ii-v). Not surprisingly, the elution order of the various Fc/2 and Fd’ glycoforms seemed to be related to the number of glycan residues involved in the formation of the biantennary/triantennary complex oligosaccharide structures (See Table S6 for details). In the reported example, the Erbitux® (Figure 4c) Fc/2 subunit presents a profile similar to that of Infliximab (Figure 2ii) and Trastuzumab (Figure 3ii), with H3N4F1 and H4N4F1 representing the most abundant moieties of the glycosylation pattern (Figures S1ii, S1iv, S1v), together with less abundant moieties such as H5N4F1 (Figure S1vi)and the afucosylated H5N2 (Figure S1iii). Interestingly, the Erbitux® Fd’ fragment was not glycosylated with such glycans.40 Instead, it was found to be populated with larger, hyper-galactosylated glycans and structures containing N-Glycolyl Neuraminic Acid (Figure S1vi-xiii); among them, H7N4F1 and H8N5F1Sg1 represent the most abundant moieties. The presence of bisecting N-acetylglucosamines forming triantennary complex oligosaccharides was also detected, with the most hydrophilic species, H9N5F1 and H8N5F1Sg1, observed to elute at the end of HILIC gradient (See Table S6 and Figure S1 for details). In the case of Cetuximab B (Figure 4d), it was possible to observe some differences versus the originator, even in the glycosylation of the Fc/2 fragment (Figure S2ii-v). For instance, it could be observed that this biosimilar differed in having an overwhelming relative abundance of H3N4F1 combined with more pronounced content of afucosylated glycans, such as H3N4, H5N2, and H4N4. Even more pronounced differences were detectable in the glycosylation of the Fd’ fragment. In contrast with Erbitux®, the Fd’ fragment of Cetuximab B (Figure S2vi-xi) was more heterogeneously populated by fucosylated glycans such as H3N3F1, H3N4F1, H4N4F1 and H5N4F1 as well as some sialylated glycans containing N-Acetyl Neuraminic Acid, of which H5N4F1Sa1 was seen to represent the most abundant moiety (See Table S6 for details).
the characterization of protein biopharmaceuticals, ranging from mAbs, next generation antibody drug-conjugates (ADCs), to other highly glycosylated bio-therapeutics such as FSH and interferons. HILIC could potentially be envisaged for routine analysis, due to its noteworthy resolving power for glyco-variants. In addition, it could also be suggested that the serial coupling and/or 2D implementation of RPLC and HILIC wide-pore columns (purposefully adapted for such configuration) could be a means to further heighten the characterization of biopharmaceuticals such that both hydrophilic and hydrophobic variants could be resolved in one unique run.
ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Schematic representation of the possible levels of LC-MS analysis and retention times and mass assignment of originators and biosimilars fragments upon middle-up RPLC- and HILIC-MS analysis (PDF)
AUTHOR INFORMATION Corresponding Author * E-mail:
[email protected] Tel.: +41 2 2379 3358 Notes The authors declare no competing financial interest.
ACKNOWLEDGMENTS Davy Guillarme wishes to thank the Swiss National Science Foundation for support through a fellow-ship to Szabolcs Fekete (31003A_159494). Jean-Luc Veuthey from the University of Geneva is acknowledged for useful comments and discussions.
REFERENCES (1) (2) (3) (4) (5) (6) (7)
CONCLUSIONS The development and approval processes for biosimilar mAbs are hinged upon their comparability to originator drug products. In this context, we show how HILIC-MS can be used as a powerful analytical tool for resolving hydrophilic variants of protein biopharmaceuticals at a middle-up level of the analysis. In particular, the direct visualization of the glycovariants allows the qualitative comparison of the glycosylation patterns between originator and biosimilar molecules. By evaluating the chromatographic profile of the HILIC middleup analysis, it is possible to quickly assess the most abundant glycosylation moieties and to highlight the differences between an originator and biosimilar. With the associated MS data, this HILIC based middle-up analysis also gives the chance to interrogate domain specific N-glycan information to an unprecedented level of detail. In our opinion, HILIC can be considered as a valid complementary approach to RPLC for
Page 6 of 8
(8) (9)
(10) (11)
(12)
(13) (14) (15) (16)
Beck, A.; Wurch, T.; Bailly, C.; Corvaia, N. Nat. Rev. Immunol. 2010, 10, 345–352. Waldmann, T. A. Nat. Med. 2003, 9, 269–277. Sharma, P.; Allison, J. P. Science (80-. ). 2015, 348, 56–61. Reichert, J. M. MAbs 2012, 4, 413–415. Reichert, J. M. MAbs 2016, 8, 197–204. Tsiftsoglou, A. S.; Ruiz, S.; Schneider, C. K. BioDrugs 2013, 27, 203–211. Thomann, M.; Reckermann, K.; Reusch, D.; Prasser, J.; Tejada, M. L. Mol. Immunol. 2016, 73, 69–75. Reusch, D.; Tejada, M. L. Glycobiology 2015, 25, 1325–1334. Alt, N.; Zhang, T. Y.; Motchnik, P.; Taticek, R.; Quarmby, V.; Schlothauer, T.; Beck, H.; Emrich, T.; Harris, R. J. Biologicals 2016, 44, 291–305. Beck, A.; Reichert, J. M. MAbs 2012, 4, 419–425. Chung, C. H.; Mirakhur, B.; Chan, E.; Le, Q.-T.; Berlin, J.; Morse, M.; Murphy, B. A.; Satinover, S. M.; Hosen, J.; Mauro, D.; Slebos, R. J.; Zhou, Q.; Gold, D.; Hatley, T.; Hicklin, D. J.; Platts-Mills, T. A. E. N. Engl. J. Med. 2008, 358, 1109–1117. Okazaki, A.; Shoji-Hosaka, E.; Nakamura, K.; Wakitani, M.; Uchida, K.; Kakita, S.; Tsumoto, K.; Kumagai, I.; Shitara, K. J. Mol. Biol. 2004, 336, 1239–1249. Hodoniczky, J.; Zheng, Y. Z.; James, D. C. Biotechnol. Prog. 2005, 21, 1644–1652. Houde, D.; Peng, Y.; Berkowitz, S. A.; Engen, J. R. Mol. Cell. Proteomics 2010, 9, 1716–1728. Nimmerjahn, F.; Anthony, R. M.; Ravetch, J. V. Proc. Natl. Acad. Sci. 2007, 104, 8433–8437. Jiang, X.-R.; Song, A.; Bergelson, S.; Arroll, T.; Parekh, B.; May, K.; Chung, S.; Strouse, R.; Mire-Sluis, A.; Schenerman,
ACS Paragon Plus Environment
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
(17)
(18)
(19) (20) (21) (22) (23) (24) (25)
(26) (27) (28) (29)
Analytical Chemistry M. Nat. Rev. Drug Discov. 2011, 10, 101–111. Goetze, A. M.; Liu, Y. D.; Zhang, Z.; Shah, B.; Lee, E.; Bondarenko, P. V.; Flynn, G. C. Glycobiology 2011, 21, 949– 959. Alessandri, L.; Ouellette, D.; Acquah, A.; Rieser, M.; LeBlond, D.; Saltarelli, M.; Radziejewski, C.; Fujimori, T.; Correia, I. MAbs 2012, 4, 509–520. Anthony, R. M.; Wermeling, F.; Karlsson, M. C. I.; Ravetch, J. V. Proc. Natl. Acad. Sci. 2008, 105, 19571–19578. Nimmerjahn, F.; Ravetch, J. V. Annu. Rev. Immunol. 2008, 26, 513–533. Jefferis, R. Nat. Biotechnol. 2006, 24, 1230–1231. Baldo, B. Oncoimmunology 2013, 2, e26333. Maeda, E.; Kita, S.; Kinoshita, M.; Urakami, K.; Hayakawa, T.; Kakehi, K. Anal. Chem. 2012, 84, 2373–2379. Toyama, A.; Nakagawa, H.; Matsuda, K.; Sato, T.; Nakamura, Y.; Ueda, K. Anal. Chem. 2012, 84, 9655–9662. Sundaram, S.; Matathia, A.; Qian, J.; Zhang, J.; Hsieh, M.-C.; Liu, T.; Crowley, R.; Parekh, B.; Zhou, Q. MAbs 2011, 3, 505– 512. Fekete, S.; Guillarme, D.; Sandra, P.; Sandra, K. Anal. Chem. 2016, 88, 480–507. Sandra, K.; Vandenheede, I.; Sandra, P. J. Chromatogr. A 2014, 1335, 81–103. Vanhoenacker, G.; Vandenheede, I.; David, F.; Sandra, P.; Sandra, K. Anal. Bioanal. Chem. 2015, 407, 355–366. Sandra, K.; Sandra, P. Bioanalysis 2015, 7, 2843–2847.
(30) (31) (32) (33) (34) (35) (36)
(37) (38) (39)
(40)
Periat, A.; Fekete, S.; Cusumano, A.; Veuthey, J.-L.; Beck, A.; Lauber, M.; Guillarme, D. J. Chromatogr. A 2016, 1448, 81–92. Chen, J.; Liu, Z.; Wang, F.; Mao, J.; Zhou, Y.; Liu, J.; Zou, H.; Zhang, Y. Chem. Commun. 2015, 51, 14758–14760. Ruta, J.; Rudaz, S.; McCalley, D. V; Veuthey, J.-L.; Guillarme, D. J. Chromatogr. A 2010, 1217, 8230–8240. Beck, A.; Reichert, J. M. MAbs 2013, 5, 621–623. Fang, J.; Doneanu, C.; Alley, W. R.; Yu, Y. Q.; Beck, A.; Chen, W. MAbs 2016, 8, 1021–1034. Sorensen, M.; Harmes, D. C.; Stoll, D. R.; Staples, G. O.; Fekete, S.; Guillarme, D.; Beck, A. MAbs 2016, 8, 1224–1234. Beck, A.; Debaene, F.; Diemer, H.; Wagner-Rousset, E.; Colas, O.; Dorsselaer, A. Van; Cianférani, S. J. Mass Spectrom. 2015, 50, 285–297. Beck, A.; Sanglier-Cianférani, S.; Van Dorsselaer, A. Anal. Chem. 2012, 84, 4637–4646. Wiegandt, A.; Meyer, B. Anal. Chem. 2014, 86, 4807–4814. Ayoub, D.; Jabs, W.; Resemann, A.; Evers, W.; Evans, C.; Main, L.; Baessmann, C.; Wagner-Rousset, E.; Suckau, D.; Beck, A. MAbs 2013, 5, 699–710. Qian, J.; Liu, T.; Yang, L.; Daus, A.; Crowley, R.; Zhou, Q. Anal. Biochem. 2007, 364, 8–18.
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 8 of 8
For TOC only
ACS Paragon Plus Environment
8
