Time Resolved Native Ion-Mobility Mass ... - ACS Publications

Article pubs.acs.org/ac

Time Resolved Native Ion-Mobility Mass Spectrometry to Monitor Dynamics of IgG4 Fab Arm Exchange and “Bispecific” Monoclonal Antibody Formation François Debaene,† Elsa Wagner-Rousset,‡ Olivier Colas,‡ Daniel Ayoub,‡ Nathalie Corvaïa,‡ Alain Van Dorsselaer,† Alain Beck,*,‡ and Sarah Cianférani*,† †

Laboratoire de Spectrométrie de Masse BioOrganique (LSMBO), IPHC-DSA, Université de Strasbourg, CNRS, UMR7178, 25 rue Becquerel, 67087 Strasbourg, France ‡ Centre d’Immunologie Pierre-Fabre (CIPF), 5 Av. Napoléon III, BP 60497, 74164 Saint-Julien-en-Genevois, France S Supporting Information *

ABSTRACT: Monoclonal antibodies (mAbs) and derivatives such as antibody−drug conjugates (ADC) and bispecific antibodies (bsAb), are the fastest growing class of human therapeutics. Most of the therapeutic antibodies currently on the market and in clinical trials are chimeric, humanized, and human immunoglobulin G1 (IgG1). An increasing number of IgG2s and IgG4s that have distinct structural and functional properties are also investigated to develop products that lack or have diminished antibody effector functions compared to IgG1. Importantly, wild type IgG4 has been shown to form half molecules (one heavy chain and one light chain) that lack interheavy chain disulfide bonds and form intrachain disulfide bonds. Moreover, IgG4 undergoes a process of Fab-arm exchange (FAE) in which the heavy chains of antibodies of different specificities can dissociate and recombine in bispecific antibodies both in vitro and in vivo. Here, native mass spectrometry (MS) and time-resolved traveling wave ion mobility MS (TWIM-MS) were used for the first time for online monitoring of FAE and bsAb formation using Hz6F4-2v3 and natalizumab, two humanized IgG4s which bind to human Junctional Adhesion Molecule-A (JAM-A) and alpha4 integrin, respectively. In addition, native MS analysis of bsAb/JAM-A immune complexes revealed that bsAb can bind up to two antigen molecules, confirming that the Hz6F4 family preferentially binds dimeric JAM-A. Our results illustrate how IM-MS can rapidly assess bsAb structural heterogeneity and be easily implemented into MS workflows for bsAb production follow up and bsAb/antigen complex characterization. Altogether, these results provide new MS-based methodologies for in-depth FAE and bsAb formation monitoring. Native MS and IM-MS will play an increasing role in next generation biopharmaceutical product characterization like bsAbs, antibody mixtures, and antibody− drug conjugates (ADC) as well as for biosimilar and biobetter antibodies.

M

IgG1 and IgG4) or four (for IgG2) disulfide bonds located in the hinge domain. The other 12 cysteine bridges are intramolecular and delimit six different globular domains: one variable (VL) and one constant for L chains and one variable (VH) and three constant for the heavy chains (CH1, CH2, and CH3). Antigen binding is mediated by the variable domains, mainly by the complementary determining regions (CDR). IgGs also present a consensus sequence for N-glycosylation in their heavy chain CH2 constant domain. The present work is focused on bsAb analytical characterization that is formed through a physiological process termed fab-arm exchange (FAE), in which half-molecules (HL pairs) of two humanized IgG4 mAbs recombine to form a chimeric bsAb protein.10,11 FAE, which naturally occurs in vivo, can be mimicked in vitro by the addition of mild reducing agents such as glutathione (Figure 1a).10 Two IgG4 residues, S228 and

onoclonal antibodies (mAbs) and related products are the fastest growing class of human therapeutics.1 More than 40 IgGs and IgG derivatives have been approved by the U.S. Food and Drug Administration (FDA)2 and the European Medicine Agency (EMA) and include “naked” antibodies, radio-immunoconjugates,3 antibody−drug conjugates (ADC),4 bispecific antibodies (bsAb),5 Fab fragments, Fc-fusion proteins/peptides,6 and immunocytokines7 for use in various indications such as cancer and inflammatory diseases,8 and 30 more are currently being investigated in phase III clinical trials.9 Most approved therapeutic mAbs are chimeric, humanized, or human IgGs with similar constant domains. IgGs are 150 kDa glycoproteins that are composed of several domains: two antigen-binding domains (Fab fragments), linked by a flexible region (the hinge) to a constant region (Fc) that interacts with immune system effector molecules. They are composed of two heavy chains (H, near 50 kDa) and two light chains (L, near 25 kDa) and form tetrameric structures (HL)2 through covalent disulfide bridges. H and L chains are linked by one disulfide bond, while two H chains are linked by two (for © 2013 American Chemical Society

Received: July 17, 2013 Accepted: September 5, 2013 Published: September 5, 2013 9785

dx.doi.org/10.1021/ac402237v | Anal. Chem. 2013, 85, 9785−9792

Analytical Chemistry

Article

Table 1. Marketed and Late Clinical Stage IgG4 mAbs (trade name)

antigen target

acute myelogenous leukemia

CD33

Nalalizumab (Tysabri)

multiple sclerosis and Crohn’s disease non-Hodgkin lymphoma eosinophilc asthma severe asthma plaque psoriasis melanoma

cell adhesion molecule α4integrin CD22

Inotuzumab ozogamicina Reslizumab Lebrikizumab Ixekizumab Nivolumab Tabalumab a

Figure 1. FAE and bsAb formation monitored by classical CEX and LC-MS analysis. (a) Schematic representation of in vitro IgG4 Fab-arm exchange mimicked by GSH reduction. (b,c) CEX profiles of equimolar mixtures of PNGase-F deglycosylated Hz6F4-2v3 and natalizumab alone (b) or after 24 h of incubation with 0.5 mM GSH at 37 °C (c). (d,e) LC-MS deconvoluted mass spectra of equimolar mixtures of PNGase-F deglycosylated Hz6F4-2v3 and natalizumab alone (d) or after 24 h of incubation with 0.5 mM GSH at 37 °C (e). Asterisks (∗) correspond to pyroglutamic acid/glutamic acid forms (pGlu/Glu) of mAbs, whereas most intense peaks correspond to pGlu/pGlu forms.

clinical indication

Gemtuzumab ozogamicina (Mylotarg)

systemic lupus erythematous

interleukin 5 interleukin 13 interleukin 17a programmed cell death 1 receptor BLysS

first approval or clinical status U. S.: 2000− 2010 Japan: 2005 US: 2004 EU: 2006 PhIII PhIII PhIII PhIII PhIII PhIII

Antibody-drug conjugate (ADC).

the promise of more effective therapeutics. The range of therapeutic uses for bsAbs is expanding beyond oncology to target diseases of the immune system, heart, and central nervous system.18,19 Recent reports and guidelines from the EMA or the FDA have pointed out that there is an increasing need for the early use of novel precise analytical techniques for mAb characterization. mAb characterization and homogeneity assessment therefore require a plethora of orthogonal analytical techniques mostly including chromatography, electrophoresis, and mass spectrometry (MS; for review, see refs 20 and 21). Recently, some structural MS-based techniques like native MS and IonMobility MS (IM-MS) have emerged for higher order structure assessment.20,22,23 Recent progress in native MS instrumentation24,25 combined with simplification of mass spectra interpretation26 has pushed native MS forward as a robust, fast, and reliable first-line ready-to-use MS approach not only for intact routine mAb analysis,22,27,28 higher order structure, and oligomeric status (dimer, trimer, tetramer) determination29 but also for mAb/Ag characterization.22,24,30 Combining ion mobility, which separates ions according to their size and shape, with MS brings an additional level of conformational characterization of protein complexes, through measurement of collision cross sections.31 Ion mobility separation is fast (millisecond measurements), sensitive (few nanomoles), and amenable to high throughput automation. Under classical MS conditions (either by direct infusion or through LC-MS experiments), IM-MS can serve for routine batch-to-batch characterization of therapeutics, for mAbs glycosylation profiling,32 for heterogeneous disulfide bridge pairing monitoring,22 or even to detect protein aggregation.33 Under native conditions, IM-MS can serve to routinely fingerprint mAbs higher order structures, to provide information on conformational changes induced upon Ag binding, or even to allow distinguishing between several isoforms.23 In a previous work, we combined a functional approach to proteomic analysis to identify new antibody target molecules. Immunization of mice with breast cancer cells (MCF-7) allowed selection of the IgG1 stabilized murine mAb mu6F4 for its remarkable in vitro and in vivo antiproliferative and antitumoral properties.34 Proteomic analysis next allowed identification of the mAb 6F4 target as the human Junctional

R409, have been identified as crucial for FAE, their mutation leading to in vitro and in vivo FAE blocking:10,12,13 S228 renders interheavy chain S−S bonds more sensitive to reduction,14 while R409 located in the CH3 domain decreases noncovalent CH3−CH3 interaction strength between half-molecules.13 In addition, the crystal structure of the human IgG4 CH3 dimer was recently reported and used to reveal the role of R409 in the FAE mechanism.15 Two humanized IgG4 antibodies, natalizumab and gemtuzumab, are approved for human use, and six more, reslizumab, lebrikizumab, ixekizumab, tabalumab, nivolumab, and inotuzumab ozogamicin, are being investigated in clinical phase III trials (Table 1).16 Natalizumab exchanges Fab arms with endogenous human IgG4 in natalizumab-treated individuals. Gemtuzumab, in contrast, contains an IgG4 core-hinge mutation that blocks Fab-arm exchange to undetectable levels both in vitro and in a mouse model.12 The ability of IgG4 therapeutics to recombine with endogenous IgG4 may affect their pharmacokinetics. Although pharmacokinetic modeling lessens concerns about undesired cross-linking under normal conditions, unpredictability remains, and mutations that completely prevent FAE in vivo should be considered when designing therapeutic IgG4 antibodies. As the FAE process occurs naturally in vivo, bsAb IgG4 cannot be used as bivalenttargeting therapeutics. Interestingly, based on in-depth structure−function relationship knowledge on IgG4 FAE, Labrijn et al. recently proposed a method for the efficient generation of stable IgG1 bsAb molecules by controlled FAE yielding bsAb drug candidates.17 Therapeutic bispecific monoclonal antibodies are thus hybrid proteins composed of fragments of two different mAbs and are of particular interest for clinical purposes as they can bind to two different types of antigen. BsAbs have a higher cytotoxic potential, which offers 9786



Article

deglycosylated by the reaction of 0.4 μL of PNGaseF (New England Biolabs, Ipswich, U. S.) per 100 μg of mAbs overnight at 37 °C. Desalting of the samples was achieved online at 0.5 mL/min with a MassPREP micro desalting cartridge 2.1 × 5 mm, 20 μm, 1000 Å (Waters, Manchester, U. K.) set at 80 °C. A short gradient was set up as follows: 2.1 min at 5% B, 0.6 min with increasing B from 5 to 90%, and a decrease to 5% B within 0.1 and 1.2 min equilibration. The mass spectrometer was operated in W positive ion mode with a capillary voltage of 3400 V and a sample cone voltage of 120 V. Acquisitions were performed in the mass range m/z 1000−4000 with a 1 s scan time. Calibration was performed using the singly charged ions produced by a 2 mg/L solution of cesium iodide in 2-propanol/ water (1/1). Data analysis was performed with a MassLynx 4.1 (Waters, Manchester, U. K.). Native Mass Spectrometry Experiments. Prior to native mass spectrometry experiments, mAbs were desalted against a 150 mM ammonium acetate solution buffered at pH 7.4 using two cycles of gel filtration columns (NAP-5, GE Healthcare, Little Chalfont, U. K.). Samples were subsequently concentrated using a centrifugal concentrator (Vivaspin, 30 kD cutoff, Sartorius, Göttingen, Germany). Antigen JAM-A was desalted using two cycles of microcentrifuge gel-filtration columns (Zeba 0.5 mL, Thermo Scientific, Rockford, IL). Protein concentration was determined with a Bradford assay using a 2 g/L BSA solution as standard. Native mass spectrometry experiments were performed on a hybrid electrospray quadrupole time-of-flight mass spectrometer (Synapt G2 HDMS, Waters, Manchester, U. K.) coupled to an automated chip-based nanoelectrospray device (Triversa Nanomate, Advion Biosciences, Ithaca, U.S.) operating in the positive ion mode. For analysis under denaturing conditions, external calibration was performed with the multiply charged ions produced by 2 μM horse heart myoglobin diluted in water/acetonitrile/formic acid (50v/50v/1v) and classical interface tuning parameters of the mass spectrometer (Vc = 40 V, Pi = 2.1 mbar). For native MS experiments, tuning parameters have been carefully optimized to improve desolvation and ion transfer as well as maintaining weak interactions. Particularly, the sample cone voltage Vc was set to 120 V and the backing pressure Pi was increased to 6 mbar to improve ion collisional cooling. In this case, external calibration was performed using singly charged ions produced by a 2 mg/mL solution of cesium iodide in 2-propanol/water (1v/1v). Native MS data interpretation was performed using a MassLynx 4.1 (Waters, Manchester, U. K.). Relative normalized species ratios of detected species presented in the Time-Resolved Native MS for FAE Monitoring section were calculated from average peak intensities of major charge states (22+ to 26+). Time-resolved native MS experiments were performed in triplicate. Traveling Wave Ion Mobility Mass Spectrometry (TWIMS). TWIMS experiments were performed using a hybrid quadrupole/ion mobility separator/time-of-flight instrument (Synapt G2 HDMS, Waters Manchester, U. K.). Ion mobility data were acquired after careful optimization of the instrumental parameters, especially by reducing the energy imparted to the ions before entering the ion mobility cell to lower ion activation while keeping sufficient desolvation. Separation parameters such as wave height (WH) and velocity (WV) were also optimized to reach the best separation resolution: Vc = 80 V, Pi = 6 mbar, Trap Bias = 40 V, Trap CE = 7 V, Transfer CE = 30 V, PHe= 180 mL/min, PN2 = 45 mL/

Adhesion Molecule-A (JAM-A). The lead murine IgG4 stabilized by a point mutation of S228 into P was then humanized (Hz6F4-2). In this first study, we used native MS and TWIMS-MS to study the formation of immune complexes involving murine or humanized 6F4 and recombinant human JAM-A. We first demonstrated that combining both MS methodologies affords powerful quality control of recombinant antigen batches in terms of structural and conformational homogeneity along with oligomerization state determination. The formation of immune complexes between JAM-A and each mAb was then investigated by native MS and showed that murine and humanized mAbs 6F4 selectively recognize JAM-A to form specific complexes containing up to four JAM-A molecules per mAb.22 In the present study, we describe the ability of native IM-MS to monitor FAE and subsequent bsAb formation using natalizumab and Hz6F4-2v3, the wild type IgG4 variant (S228) of our lead candidate Hz6F4-2. In addition, we monitor bsAb/antigen complex formation and determine its stoichiometry using native MS.

■

EXPERIMENTAL SECTION Production and Purification of Recombinant Antigen and Antibodies. JAM-A production and purification was performed as described by Atmanene et al.22 Natalizumab was produced by Biogen Idec in NS0 cells.35 Recombinant antibodies (Hz6F4-2 and Hz6F4-2v3) used in this study were expressed in eukaryotic cell lines36 at the Centre d’Immunologie Pierre Fabre (Saint-Julien-en-Genevois, France) and by PX’Therapeutics (Grenoble, France) and purified using standard manufacturing procedures.37 Sample Preparation for bsAb Formation. Hz6F4-2v3 was diluted and mixed either with natalizumab (IgG4), Hz6F42 (IgG1), or trastuzumab (IgG1) at a final concentration of 0.75 mg/mL for each antibody with 1× PBS in a final volume of 50 μL. Mixtures of pairs of antibodies were incubated 24 h at 37 °C. GSH mediated FAE reaction occurred in the presence of 0.5 mM reduced glutathione (GSH, Sigma). Cation Exchange Chromatography. CEX experiments were performed using an HPLC Waters 2695 equipped with a 2487 UV detector (Waters, Manchester, U. K.). Separation of natalizumab, Hz6F4-2v3, Hz6F4-2, and bsAb was performed with a Propac WCX-10 4 × 250 mm (Thermo Scientific, Rockford, IL) with a gradient elution (mobile phase A, 20 mM sodium phosphate, pH 5.5; mobile phase B, 20 mM sodium phosphate, 1 M NaCl, pH 6.5; gradient, 4−15% B in 73 min at a flow rate of 1 mL/min). Samples were diluted twice in A before injection, and 15 μg of each mAb were loaded on the column. Bispecific Antibody Purification by CEX. Larger amounts of bsAb were prepared as described previously from 500 μg of each antibody as starting material. After incubation, the mix solution was concentrated using a centrifugal concentrator (Vivaspin, 30 kDa, Sartorius, Göttingen, Germany) to a final volume of 150 μL. For each collection, 100 μL of the sample diluted twice in CEX mobile phase A was loaded on a Propac WCX-10 4 × 250 mm column (Thermo Scientific, Rockford, IL). Collected fractions of bsAb were pooled and concentrated, resulting in a 0.8 mg/mL bsAb solution. LC-MS Experiments. LC-MS experiments were performed using an ultrahigh-performance liquid chromatography system (UHPLC Acquity, Waters, Manchester, U. K.) coupled to an electrospray time-of-flight mass spectrometer (LCT Premier, Waters, Manchester, U. K.). Antibody mixtures were 9787



Article

min, IMS WV = 830 m/s, IMS WH = 31 V. Reported collision cross sections correspond to the average CCS measured for the 24+ to 27+ charge states of each mAb, using alcohol deshydrogenase (24+ to 26+ charge states) as an external calibrant.38,39 IM-MS measurements have been performed in triplicate, and data interpretation has been performed using MassLynx 4.1 and Driftscope version 2.2 (Waters, Manchester, U. K.). Fab Arm Exchange Reaction. Natalizumab and Hz6F42v3 were buffer exchanged through two cycles in gel filtration columns (NAP-5, GE Healthcare, Little Chalfont, U. K.) against 150 mM ammonium acetate, pH 7.4. Both mAb solutions were diluted to 1 mg/mL in 150 mM ammonium acetate, pH 7.4, and further incubated with 0.5 mM glutathione (GSH) at 37 °C to reach a final concentration of 0.5 mg/mL. In order to remove excess GSH, 10 μL of the mixture obtained after different incubation time points (5 min, 30 min, 1 h, 2 h, 4 h, 6 h, 8 h, 18 h, and 24 h) were desalted against 150 mM ammonium acetate, pH 7.4, through four cycles of dilution/ concentration using a centrifugal concentrator (Vivaspin, 30 kD, Sartorius, Göttingen, Germany) and further analyzed with native MS and IM-MS.

Figure 2. Denaturing and native ESI-MS analysis of the purified natalizumab/Hz6F4-2v3 bsAb. ESI mass spectra of the purified natalizumab/Hz6F4-2v3 bispecific antibody obtained in classical denaturing (H2O/ACN/FA 50:50:1) (a) and native (AcONH4 150 mM, pH 7.4) (b) conditions. Insets present a zoom of the 45+ and 24+ charge states of natalizumab/Hz6F4-2v3 bsAb under denaturing and native conditions, respectively. (c) Table summarizing measured and theoretical molecular masses of mAbs G0F/G1F main glycoform.

■

RESULTS AND DISCUSSION FAE and bsAb Formation Monitored by Classical CEX and LC-MS Analysis. FAE was mimicked in vitro by the addition of glutathione (GSH) as a mild reducing agent to the two deglycosylated IgG4 mAbs, natalizumab and Hz6F4-2v3 (see Experimental Section). We first monitored natalizumab/ Hz6F4-2v3 bsAb formation by cation exchange chromatography (CEX) and classical liquid chromatography coupled to MS analysis (LC-MS). After 24 h of incubation with 0.5 mM GSH at 37 °C, an additional CEX peak is observed with a retention time intermediate (32 min) to individual natalizumab (18 min) and Hz6F4-2v3 (47 min) mAbs (Figure 1b,c). A control experiment involving an IgG1 mAb (trastuzumab) and the Hz6F4-2v3 IgG4 (S228, wild type) was performed in parallel, and no additional peak was observed (see Supporting Information (SI) Figure S1). As the FAE phenomenon induces spontaneous interchain disulfide bond reformation, bsAb formation was next analyzed by LC-MS analysis. As depicted in Figure 1d,e, an additional ion series with a molecular mass of 145246 ± 1 Da, corresponding to the aglycosylated natalizumab/Hz6F4-2v3 bsAb (theoretical MW is 145245 Da), is observed after GSH treatment. Conversely, no additional species was detected after the reaction involving trastuzumab and Hz6F4-2v3 (see Supporting Information (SI) Figure S1). Native MS and IM-MS Analysis of Purified Natalizumab/Hz6F4-2v3 bsAb. Larger amounts of natalizumab/ Hz6F4-2v3 bsAb were obtained from CEX purification (see Experimental Section). Purified glycosylated bsAb, as well as both starting IgG4’s (see Supporting Information (SI) Figure S2), were further subjected to denaturing and native MS analysis. Under classical denaturing conditions, as previously observed in LC-MS analysis, a main ion series corresponding to the characteristic glycosylation profile of natalizumab/Hz6F42v3 bsAb is observed (Figure 2a and c), suggesting that a high proportion of disulfide bound between half-mAbs recombines after CEX purification. In addition, minor ion series with molecular mass of ∼73 kDa were also observed and correspond to small amounts of natalizumab and Hz6F4-2v3 half molecules (HL pairs), respectively. In addition, CEX, nonreducing CE-

SDS, and MS experiments showed that about 80% of disulfide bonds are formed (see Supporting Information (SI) Figures S2 and S3), which supports HL pair detection in our denaturing MS experiment. These observations are in agreement with those reported by Rispens et al., who found that under mild reducing conditions (e.g., 0.5 mM GSH), only a modest fraction (∼12%) of IgG4 does not contain covalent bonds between H chains.40 Native MS analysis (Figure 2b and c) of the purified bsAb reveals an almost exclusive homogeneous ion population in the m/z range 5000−7500, corresponding to natalizumab/Hz6F4-2v3 bsAb with a measured MW of 148 300 ± 3 Da for the G0F/G1F glycoform. As expected, a smaller proportion of half antibodies is detected in native MS in the lower m/z range than under classical denaturing conditions, which can be explained by residual noncovalent interactions involved in bsAb formation by FAE. We next analyzed the purified natalizumab/Hz6F4-2v3 bsAb using native IM-MS (Figure 3). As shown on the driftscope of the purified bsAb (Figure 3a), carefully tuned ion mobility cell parameters led to an almost complete separation of all charge states, corresponding to a resolving power tD/ΔtD at FWHM of a Synapt G2 ion mobility cell of 12.8. Figure 3b presents IM arrival time distributions (ATD) obtained for the 24+ charge state of natalizumab, Hz6F4-2v3, and natalizumab/Hz6F4-2v3 bsAb. As expected, the IM-MS drift time (tD) observed for the newly formed natalizumab/Hz6F4-2v3 antibody is intermediate to individual starting mAbs tD’s. Collision cross sections (CCS) were calculated on the basis of the four most intense charge states of each mAb (24+ to 27+). All three mAbs revealed very close gas phase conformations: natalizumab presents the most extended conformation with a CCS of 69.6 ± 0.4 nm2, Hz6F42v3 being the lowest with a CCS of 65.6 ± 0.3 nm2, and the bispecific antibody with an intermediate CCS of 66.9 ± 0.1 nm2 (Figure 3c). The central point of our work is in fact the ability of currently available IM-MS instrumentation to differentiate 9788



Article

native MS in order to address the stoichiometry question, binding of two JAM-A molecules being expected for the bsAb. A fixed amount of natalizumab/Hz6F4-2v3 bsAb (5 μM, Figure 4a) was incubated with increasing amounts of JAM-A. As

Figure 4. Determination of bsAb/Ag binding stoichiometries. Deconvoluted native ESI mass spectra of (a−c) natalizumab/Hz6F42v3 bsAb (5 μM) alone (a) and in the presence of 10 μM (b) or 40 μM (c) JAM-A. (d) Hz6F4-2v3 (5 μM) in the presence of 40 μM JAM-A. MWs correspond to the main G0F/G1F glycoform.

Figure 3. Native IM-MS analysis of purified natalizumab/Hz6F4-2v3 bsAb. (a) Driftscope of the purified natalizumab/Hz6F4-2v3 bsAb. (b) ATDs corresponding to the 24+ charge state of natalizumab (black), purified natalizumab/Hz6F4-2v3 bsAb (red), and Hz6F4-2v3 (blue). (c) Table summarizing IM-MS calculated CCS for each mAb. CCS standard deviations were obtained from IM-MS measurements performed in triplicate.

shown in Figure 4b, when bispecific mAb and JAM-A are present in equimolar concentrations, two species are detected: the intact bispecific mAb (MW 148 302 ± 1 Da) and one additional compound corresponding to the formation of a 1:1 bsAb/Ag complex (MW 172 369 ± 3 Da). Then, increasing the JAM-A concentration (up to eight molar equivalents, 40 μM) induces an almost quantitative shift of the equilibrium toward the formation of the 1:2 bispecific bsAb/Ag complex (Figure 4c). Under similar conditions (40 μM JAM-A), the parent Hz6F4-2v3 mAb is able to bind up to four JAM-A molecules (Figure 4d) similarly to the lead Hz6F4-2 construct.22 Altogether, these native MS analyses allowed the conclusion that bsAb natalizumab/Hz6F4-2v3 is able to bind up to two antigen molecules, reinforcing the hypothesis that JAM-A preferentially binds to Hz6F4-2 as a dimer. Time-Resolved Native MS and IM-MS Online Monitoring of FAE with Natalizumab and Hz6F4-2v3. TimeResolved Native MS for FAE Monitoring. Native MS provides a reliable method for qualitative and semiquantitative analysis of mixtures of antibodies, including mixtures of mono- and bispecific antibodies.27 Rose et al. have reported the use of native MS to measure the equilibrium constants of IgG4 dissociation into half molecules, as well as the kinetics involved, providing insight into the importance of the CH3 domain in the antibody for FAE.41,42 In a similar way, we used here time-

between former IgG4 mAbs and the newly formed bsAb. As no 3D crystal structure is available for any of our mAbs, the question of the true structural difference between the studied species is legitimate. In our experience, the mass difference between compounds cannot be correlated to a corresponding CCS difference. Thus, even if mass difference between bsAb and constitutive IgG4 has an impact on IM drift time and CCS measurement, we believe that our observations reflect a real structural difference. Binding Stoichiometries of Bispecific/Antigen Immune Complexes: Natalizumab/Hz6F4-2v3 bsAb Binds up to Two JAM-A Molecules. Native mass spectrometry has recently been shown to be a powerful technique to determine mAb/Ag binding stoichiometry.22,24,30 In our previous work, native MS was the only methodology to unambiguously determine Hz6F4-2/antigen binding stoichiometry, revealing that the lead hinge stabilized Hz6F4-2 is able to bind up to four JAM-A molecules.22 These data strongly suggested that Hz6F42 preferentially recognizes the dimeric form of JAM-A. To go ahead, titration experiments involving bispecific natalizumab/Hz6F4-2v3 were conducted and analyzed by 9789



Article

constant for the natalizumab/Hz6F4-2v3 bsAb formation is about 4.66 × 10−5 s−1, which is in agreement with previously observed kinetics of FAE (7−10 × 10−5 s−1 as reported in refs 40 and 41). Time-Resolved Native IM-MS for FAE Monitoring. We next evaluated the possibilities of IM-MS to monitor the formation of bsAb as a function of time. Figure 6a presents driftscope

resolved native MS to monitor FAE kinetics involving natalizumab and Hz6F4-2v3 mAbs without prior deglycosylation. Native mass spectra obtained at different time points after GSH treatments are shown in Figure 5a−c. The upper mass

Figure 6. Time-resolved native IM-MS for FAE monitoring. (a−c) Driftscopes of FAE reaction between natalizumab and Hz6F4-2v3 at 0 min (a), after 8 h of reaction (b), and in the absence of GSH after 8 h (c). (d) ATDs corresponding to the 24+ charge state of G0F/G1F glycoform of natalizumab (black), purified natalizumab/Hz6F4-2v3 bsAb (red), and Hz6F4-2v3 (blue).

Figure 5. Time-resolved native MS for FAE monitoring. (a−c) nanoESI native mass spectra of the FAE reaction between natalizumab and Hz6F4-2v3 at 0 min (a), after 4 h (b), and after 24 h (c). On the right side, a zoom of the 24+ charge state is presented. (d) Graphic showing relative amounts of natalizumab/Hz6F4-2v3 bsAb (red), starting natalizumab (black), and Hz6F4-2v3 (blue) quantified at different time points from native mass spectra.

plots of mixtures of natalizumab and Hz6F4-2v3 in the presence of GSH at different time points. At t = 0 min, individual mAbs are separated in the IMS cell, and two different populations are clearly observed on the driftscope for both 23+ and 24+ charge states (Figure 6a). After 8 h of GSH treatment, a third population with an intermediate drift time is observed on the driftscope, corresponding to the formation of the natalizumab/Hz6F4-2v3 bsAb (Figure 6b). In the absence of GSH, no additional species with intermediate tD is observed (Figure 6c). ATDs corresponding to the 24+ charge state of G0F/G1F ions of all three species are depicted in Figure 6d. All ATDs are symmetric, and significantly different tD’s are observed for natalizumab (tD = 21.3 ms), Hz6F4-2v3 (tD = 20.4 ms), and natalizumab/Hz6F4-2v3 bsAb (tD = 20.8 ms), which is consistent with tD’s observed for individual species (Figure 3b). It could be noticed that ATDs in the presence of GSH are broader than those for individual purified mAbs, leading to a subsequent slight shift in tD values. A possible explanation for this observation might be that traces of GSH are still present even after extensive desalting, which affects nanospray properties, therefore resulting in less efficient desolvation reflected by slightly broader ATD peak shapes. While native MS has already been used in a few cases to follow up dynamics and kinetics of noncovalent complex formation,41,43,45,46 to our knowledge it is the first time that IMMS is used to monitor reaction dynamics. One key issue in IMMS experiments is IM resolution. While second generation T-

spectrum presents both starting mAbs before GSH addition (Figure 5a). After 4 h of reaction, first traces of bsAb are detected with an additional ion series corresponding to the glycosylation profile of natalizumab/Hz6F4-2v3 bsAb (Figure 5b and 5d). After 24 h, equilibrium is reached, and complete exchange (50% mixed bispecific (HL)2) occurs (Figure 5c and d). Control experiments performed either without GSH or in the presence of GSH and a mixture of the IgG1 trastuzumab and IgG4 natalizumab (see Supporting Information (SI) Figure S4) do not lead to the detection of any bsAb, which unambiguously assess the specificity of the formed complexes. As native MS enables the detection of all species simultaneously present in solution and because native MS is a relatively fast analytical technique (data acquired within a few minutes), the formation of mixed bispecific (HL)2 species could be monitored as a function of time (Figure 5d). Apparent rate constants were derived from these data using a simplified firstorder exponential fit.43,44 Although this kinetic model is an oversimplification of the system, the order of magnitude of the rate constant gives a good indication of the kinetic behavior of the Hz6F4-2v3 construct. The apparent pseudo-first-order rate 9790



Article

Wave cells available on Synapt G2 HDMS instruments47 do not have enough resolution to differentiate mAb glycoforms with mass differences of 140−160 Da, we demonstrate here that mixtures of glycosylated mAbs with very close MWs differing by 600−700 Da (∼0.5% of the total molecular mass) can be differentiated by IM-MS. IM separation of conformers displaying only a subtle CCS difference requires fine-tuned instrument settings to achieve maximal IM resolution. We therefore optimized IM separation parameters by adjusting gas flow rates and wave parameters (wave height and velocity) applied to the IM region of the Synapt G2 HDMS to reach a maximum resolution factor (Rs, see Atmanene et al. for Rs definition38) of 0.25 between two species.

IM cells, IM-MS can serve to detect species with very close MWs. Altogether, these results provide new MS-based methodologies for in-depth FAE and bsAb formation monitoring. Clearly native MS and IM-MS will play a central role in next generation complex biopharmaceutical product characterization such as bsAbs, antibody mixtures, antibody−drug conjugates (ADC), and obviously also for biosimilar and biobetter antibody and Fc-fusion proteins.

■

ASSOCIATED CONTENT

S Supporting Information *

(Figure S1) Control experiments of FAE and bsAb formation monitored by classical CEX and LC-MS analysis with trastuzumab (IgG1) and Hz6F4-2v3 (IgG4). (Figure S2) Denaturing and native ESI-MS analysis of individual natalizumab and Hz6F4-2v3 mAbs. (Figure S3) CE-SDS (UV 220 nm) profiles of Hz6F4-2 and Hz6F4-2v3. (Figure S4) Time-resolved native MS for FAE monitoring. This material is available free of charge via the Internet at http://pubs.acs.org.

■

CONCLUSIONS IgG4 bsAbs are formed through a physiological process termed Fab-arm exchange (FAE), in which half-molecules of two wild type IgG4 mAbs recombine to form heterotetramers with two different light and heavy chains. In the present study, we have described the combination of native MS to IM-MS for online monitoring of FAE and bsAb formation using a humanized IgG4. Particularly, we describe for the first time, a time-resolved IM-MS experiment that allowed us to monitor in real time FAE and the subsequent bispecific mAbs formation. In addition, native MS analysis of bsAb/JAM-A immune complexes revealed that bsAb can bind up to two antigen molecules, confirming that Hz6F4-2 preferentially binds dimeric JAM-A. Our results illustrate how IM-MS can rapidly assess bsAb structural heterogeneity and be easily implemented into MS workflows for bsAb production follow up and bsAb/antigen complex characterization. Among emergent MS techniques, native MS has now reached a level where it will be used in biopharmaceutical companies for antibody lead selection, optimization, and quality control. Except sample preparation (desalting), which is still hard to automatize because it could be different from one mAb to another one, native MS is now ready to use for high throughput routine mAb analysis (from data acquisition to data treatment). Originally performed on TOF or Q-TOF instruments, recent instrumental improvements allow a benefit from high resolution on FTICR48 and Orbitrap24 instruments, for example. Ultra-high resolution will become increasingly important for mAb analysis, as it would provide isotopic analysis and subsequent primary sequence assessment within one unique intact mAb analysis. A modified Orbitrap instrument adapted for native MS analysis has recently been described, opening new doors for gaining high resolution under native conditions.24 This will be particularly interesting to improve quantitative native MS data interpretation. Native IM-MS developments are not so far along for the moment. However, it is of interest for quality control of mAbs, as already depicted.21,29 Even if IgG domain flexibility makes IM-MS a challenging technique for absolute deduction of mAb structures, it can easily be incorporated into MS workflows for comparability studies, purity control, and aggregation assessment. The present study illustrates how IM-MS can rapidly assess structural variability of antibodies and monitor FAE processes. In the present paper, the potential of IM-MS was first described on individual purified species. We next describe for the first time time-resolved IM-MS experiments that allowed us real time monitoring of FAE involving two IgG4 mAbs and the subsequent bispecific mAbs formation. Despite the quite low resolving power of current commercially available

■

AUTHOR INFORMATION

Corresponding Authors

*Phone: +33 (0)4 50 35 35 22. Fax: +33 (0)4 50 35 35 90. Email: [email protected]. *Phone: +33 (0)3 68 85 26 79. Fax: +33 (0)3 68 85 27 81. Email: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge L. Morel-Chevillet and M. Excoffier (CIPF) for their contribution in chromatographic and electrophoretic characterization of mAbs. This work was supported by the OptimAbs network bioclusters (LyonBiopole and Alsace Biovalley) and sponsors (DGCIS, Oséo, Feder, Régions RhôneAlpes and Alsace, Communauté Urbaine de Strasbourg, CNRS, University of Strasbourg). We thank the GIS IBiSA for financial support of a Synapt G2 HDMS mass spectrometer.

■

ABBREVIATIONS mAb monoclonal antibody bsAb bispecific antibody Ag antigen IgG immunoglobulin G ESI electrospray ionization TWIMS traveling wave ion mobility mass spectrometry IM-MS ion mobility mass spectrometry FAE Fab-arm exchange JAM-A Junctional Adhesion Molecule-A H heavy chain L light chain

■

REFERENCES

(1) Beck, A.; Wurch, T.; Bailly, C.; Corvaia, N. Nat. Rev. Immunol. 2010, 10, 345−352. (2) Swann, P. G.; Tolnay, M.; Muthukkumar, S.; Shapiro, M. A.; Rellahan, B. L.; Clouse, K. A. Curr. Opin. Immunol. 2008, 20, 493−499. (3) Steiner, M.; Neri, D. Clin. Cancer Res. 2011, 17, 6406−6416. (4) Beck, A.; Haeuw, J. F.; Wurch, T.; Goetsch, L.; Bailly, C.; Corvaia, N. Discovery Med. 2010, 10, 329−339. (5) Beck, A.; Reichert, J. M. MAbs 2011, 3, 221−222. (6) Beck, A.; Reichert, J. M. MAbs 2011, 3, 415−416. 9791



Article

(7) Pasche, N.; Neri, D. Drug DiscoveryToday. 2012, 17, 583−590. (8) Reichert, J. M. MAbs 2012, 4, 413−415. (9) Reichert, J. M. MAbs 2012, 4, 1−3. (10) van der Neut Kolfschoten, M.; Schuurman, J.; Losen, M.; Bleeker, W. K.; Martínez-Martínez, P.; Vermeulen, E.; den Bleker, T. H.; Wiegman, L.; Vink, T.; Aarden, L. A.; De Baets, M. H.; van de Winkel, J. G. J.; Aalberse, R. C.; Parren, P. W. H. I. Science 2007, 317, 1554−1557. (11) Schuurman, J.; Labrijn, A. F.; Parren, P. W. MAbs 2012, 4, 636. (12) Labrijn, A. F.; Buijsse, A. O.; van den Bremer, E. T.; Verwilligen, A. Y.; Bleeker, W. K.; Thorpe, S. J.; Killestein, J.; Polman, C. H.; Aalberse, R. C.; Schuurman, J.; van de Winkel, J. G.; Parren, P. W. Nat. Biotechnol. 2009, 27, 767−71. (13) Labrijn, A. F.; Rispens, T.; Meesters, J.; Rose, R. J.; den Bleker, T. H.; Loverix, S.; van den Bremer, E. T.; Neijssen, J.; Vink, T.; Lasters, I.; Aalberse, R. C.; Heck, A. J.; van de Winkel, J. G.; Schuurman, J.; Parren, P. W. J. Immunol. 2011, 187, 3238−46. (14) Bloom, J. W.; Madanat, M. S.; Marriott, D.; Wong, T.; Chan, S. Y. Protein Sci. 1997, 6, 407−15. (15) Davies, A. M.; Rispens, T.; den Bleker, T. H.; McDonnell, J. M.; Gould, H. J.; Aalberse, R. C.; Sutton, B. J. Mol. Immunol. 2013, 54, 1− 7. (16) Reichert, J. M. MAbs 2013, 5, 1−4. (17) Labrijn, A. F.; Meesters, J. I.; de Goeij, B. E.; van den Bremer, E. T.; Neijssen, J.; van Kampen, M. D.; Strumane, K.; Verploegen, S.; Kundu, A.; Gramer, M. J.; van Berkel, P. H.; van de Winkel, J. G.; Schuurman, J.; Parren, P. W. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 5145−50. (18) Beck, A.; Carter, P. J.; Gerber, H. P.; Lugovskoy, A. A.; Wurch, T.; Junutula, J. R.; Kontermann, R.; Mabry, R.; Ghayur, T. MAbs 2013, 5. (19) Kontermann, R. MAbs 2012, 4. (20) Beck, A.; Sanglier-Cianferani, S.; Van Dorsselaer, A. Anal. Chem. 2012, 84, 4637−46. (21) Beck, A.; Wagner-Rousset, E.; Ayoub, D.; Van Dorsselaer, A.; Sanglier-Cianferani, S. Anal. Chem. 2013, 85, 715−36. (22) Atmanene, C.; Wagner-Rousset, E.; Malissard, M.; Chol, B.; Robert, A.; Corvaia, N.; Van Dorsselaer, A.; Beck, A.; SanglierCianferani, S. Anal. Chem. 2009, 81, 6364−73. (23) Bagal, D.; Valliere-Douglass, J. F.; Balland, A.; Schnier, P. D. Anal. Chem. 2010, 82, 6751−6755. (24) Rosati, S.; Rose, R. J.; Thompson, N. J.; van Duijn, E.; Damoc, E.; Denisov, E.; Makarov, A.; Heck, A. J. Angew. Chem., Int. Ed. 2012, 51, 12992−6. (25) Rose, R. J.; Damoc, E.; Denisov, E.; Makarov, A.; Heck, A. J. Nat. Methods 2012, 9, 1084−6. (26) Valliere-Douglass, J. F.; McFee, W. A.; Salas-Solano, O. Anal. Chem. 2012, 84, 2843−2849. (27) Rosati, S.; Thompson, N. J.; Barendregt, A.; Hendriks, L. J.; Bakker, A. B.; de Kruif, J.; Throsby, M.; van Duijn, E.; Heck, A. J. Anal. Chem. 2012, 84, 7227−32. (28) Thompson, N. J.; Rosati, S.; Rose, R. J.; Heck, A. J. Chem Commun (Cambridge, U.K.) 2013, 49, 538−48. (29) Kukrer, B.; Filipe, V.; van Duijn, E.; Kasper, P. T.; Vreeken, R. J.; Heck, A. J.; Jiskoot, W. Pharm. Res. 2010, 27, 2197−2204. (30) Atmanene, C.; Wagner-Rousset, E.; Corvaia, N.; Van Dorsselaer, A.; Beck, A.; Sanglier-Cianferani, S. Methods Mol. Biol. 2013, 988, 243−68. (31) Thalassinos, K.; Slade, S. E.; Jennings, K. R.; Scrivens, J. H.; Giles, K.; Wildgoose, J.; Hoyes, J.; Bateman, R. H.; Bowers, M. T. Int. J. Mass Spectrom. 2004, 236, 55−63. (32) Olivova, P.; Chen, W.; Chakraborty, A. B.; Gebler, J. C. Rapid Commun. Mass Spectrom. 2008, 22, 29−40. (33) Woods, L. A.; Radford, S. E.; Ashcroft, A. E. Biochim. Biophys. Acta 2013, 1834, 1257−68. (34) Goetsch, L.; Haeuw, J. F.; Beau-Larvor, C.; Gonzalez, A.; Zanna, L.; Malissard, M.; Lepecquet, A. M.; Robert, A.; Bailly, C.; Broussas, M.; Corvaia, N. Int. J. Cancer 2013, 132, 1463−74.

(35) Shapiro, R. I.; Plavina, T.; Schlain, B. R.; Pepinsky, R. B.; Garber, E. A.; Jarpe, M.; Hochman, P. S.; Wehner, N. G.; Bard, F.; Motter, R.; Yednock, T. A.; Taylor, F. R. J. Pharm. Biomed. Anal. 2011, 55, 168− 75. (36) Chartrain, M.; Chu, L. Curr. Pharm. Biotechnol. 2008, 9, 447−67. (37) Beck, A.; Bussat, M. C.; Zorn, N.; Robillard, V.; KlinguerHamour, C.; Chenu, S.; Goetsch, L.; Corvaia, N.; Van Dorsselaer, A.; Haeuw, J. F. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2005, 819, 203−218. (38) Atmanene, C.; Petiot-Becard, S.; Zeyer, D.; Van Dorsselaer, A.; Vivat Hannah, V.; Sanglier-Cianferani, S. Anal. Chem. 2012, 84, 4703− 10. (39) Bush, M. F.; Hall, Z.; Giles, K.; Hoyes, J.; Robinson, C. V.; Ruotolo, B. T. Anal. Chem. 2010, 82, 9557−65. (40) Rispens, T.; Ooijevaar-de Heer, P.; Bende, O.; Aalberse, R. C. J. Am. Chem. Soc. 2011, 133, 10302−11. (41) Rose, R. J.; Labrijn, A. F.; van den Bremer, E. T.; Loverix, S.; Lasters, I.; van Berkel, P. H.; van de Winkel, J. G.; Schuurman, J.; Parren, P. W.; Heck, A. J. Structure 2011, 19, 1274−82. (42) Rose, R. J.; van Berkel, P. H.; van den Bremer, E. T.; Labrijn, A. F.; Vink, T.; Schuurman, J.; Heck, A. J.; Parren, P. W. MAbs 2013, 5, 219−28. (43) Chevreux, G.; Atmanene, C.; Lopez, P.; Ouazzani, J.; Van Dorsselaer, A.; Badet, B.; Badet-Denisot, M. A.; Sanglier-Cianferani, S. J. Am. Soc. Mass Spectrom. 2011, 22, 431−9. (44) Rose, R. J.; Verger, D.; Daviter, T.; Remaut, H.; Paci, E.; Waksman, G.; Ashcroft, A. E.; Radford, S. E. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 12873−8. (45) Deroo, S.; Hyung, S. J.; Marcoux, J.; Gordiyenko, Y.; Koripella, R. K.; Sanyal, S.; Robinson, C. V. ACS Chem. Biol. 2012, 7, 1120−7. (46) Hyung, S. J.; Deroo, S.; Robinson, C. V. ACS Chem. Biol. 2010, 5, 1137−46. (47) Giles, K.; Williams, J. P.; Campuzano, I. Rapid Commun. Mass Spectrom. 2011, 25, 1559−66. Zhong, Y.; Hyung, S. J.; Ruotolo, B. T. Analyst 2011, 136, 3534−41. (48) Fornelli, L.; Damoc, E.; Thomas, P. M.; Kelleher, N. L.; Aizikov, K.; Denisov, E.; Makarov, A.; Tsybin, Y. O. Mol. Cell. Proteomics 2012, 11, 1758−67.

9792


Time Resolved Native Ion-Mobility Mass ... - ACS Publications

Recommend Documents