Extensive Chemical Modifications in the Primary Protein Structure of

Subscriber access provided by University of Newcastle, Australia

Brief Article

Extensive chemical modifications in the primary protein structure of IgG1 subvisible particles are necessary for breaking immune tolerance Bjoern Boll, Juliana Bessa, Emilien Folzer, Anacelia Ríos Quiroz, Roland Schmidt, Patrick Bulau, Christof Finkler, Hanns-Christian Mahler, Jörg Huwyler, Antonio Iglesias, and Atanas V. Koulov Mol. Pharmaceutics, Just Accepted Manuscript • DOI: 10.1021/acs.molpharmaceut.6b00816 • Publication Date (Web): 16 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.

Molecular Pharmaceutics 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 22

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

Molecular Pharmaceutics

Extensive chemical modifications in the primary protein structure of IgG1 subvisible particles are necessary for breaking immune tolerance Björn Boll1,5, Juliana Bessa2, Emilien Folzer1,3,5, Anacelia Ríos Quiroz1,5#, Roland Schmidt3, Patrick Bulau4, Christof Finkler1, Hanns-Christian Mahler3, Jörg Huwyler5, Antonio Iglesias2, Atanas V. Koulov1

1

Analytical Development & Quality Control, Pharma Technical Development Biologics Europe, F. Hoffmann-La Roche Ltd., Basel, Switzerland 2 Roche Pharmaceutical Research and Early Development, Pharmaceutical Sciences, Roche Innovation Center, Basel, Switzerland 3 Pharmaceutical Development & Supplies, Pharma Technical Development Biologics Europe, F. Hoffmann-La Roche Ltd, Basel, Switzerland 4 Analytical Development & Quality Control, Pharma Technical Development Biologics Europe, Roche Diagnostics GmbH, Penzberg, Germany 5 Division of Pharmaceutical Technology, Department of Pharmaceutical Sciences, University of Basel, Basel, Switzerland #

Corresponding author: Anacelia Ríos Quiroz, Email: [email protected]

Current affiliations: Hanns-Christian Mahler Drug Product Services, Lonza AG, CH - 4002 Basel e-mail:[email protected] Atanas V. Koulov Analytical Development & Quality Control, Drug Product Services, Lonza AG, CH - 4002 Basel e-mail:[email protected] Björn Boll Biologics Technical Development & Manufacturing, Novartis Pharma AG, CH-4002 Basel, Switzerland e-mail: [email protected]

1 ACS Paragon Plus Environment


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 22

Abstract A current concern with the use of therapeutic proteins is the likely presence of aggregates and submicron, subvisible and visible particles. It has been proposed that aggregates and particles may lead to unwanted increases in immune response with a possible impact on safety or efficacy. The aim of this study was thus to evaluate the ability of subvisible particles of a therapeutic antibody to break immune tolerance in an IgG1 transgenic mouse model and to understand the particle attributes that may play a role in this process. We investigated the immunogenic properties of subvisible particles (unfractionated, mixed populations and well-defined particle size-fractions) using a transgenic mouse model expressing a mini-repertoire of human IgG1 (hIgG1 tg). Immunization with proteinaceous subvisible particles generated by artificial stress conditions demonstrated that only subvisible particles bearing very extensive chemical modifications within the primary amino acid structure were able to break immune tolerance in the hIgG1 transgenic mouse model. Protein particles exhibiting low levels of chemical modification were not immunogenic in this model. Key Words: subvisible protein particles - chemical modifications - immunogenicity IgG1 tg mouse - biotherapeutics

For TOC use:


Page 3 of 22

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


Introduction Comprehensive formulation development, primary packaging design and process design is needed and performed aiming to minimize levels of particulates for a given parenteral drug product. However, all parenteral products contain particulates of some size and kind. Particles can be proteinaceous or of other nature and composition, e.g. glass, cellulose fibers, silicone droplets, etc.(1, 2) Protein aggregation is an intrinsic and ubiquitous property of all proteins. Consequently, particles and aggregates are present to some extent in all biotherapeutic formulations used in clinical trials and for marketed drugs.(3-8) Such particles of proteinaceous origin have recently received increased interest as result of industry, academia and regulators’ hypothesis that it could lead to possible biological consequences such as immunogenicity, and/or altered bioactivity and pharmacokinetics.(9-11) Immunogenic effects after drug administration result in the formation of anti-drug antibodies (ADAs) specific for the protein of interest.(12) The undesired immunogenicity of a protein product might result in reduced efficacy of the drug and/or lead to safety-related adverse events. The current knowledge of the relationship between immunogenicity and protein particles is based on studies of vaccines and the immunogenicity of protein assemblies.(13-15) Rigidly packed arrays of foreign proteins in the micrometer range have been shown to be more immunogenic then their monomeric counterparts. However, despite the theoretical concerns that allogeneic proteinaceous particles could be a potential risk factor for breaking immune tolerance, their biological effects are still unclear. Furthermore, a definite link between the intrinsic immunogenicity of human or humanized protein therapeutics in the clinics and sub-visible particle content in biotherapeutics is lacking. To date, the available data from in vitro and in vivo experiments has been controversial and incomplete, which makes it impossible to reach clear conclusions. Two major caveats of the studies published to date are: a) the use of insufficiently characterized species or b) the use of complex mixtures of monomers, various aggregates and broadly-distributed particle populations. (13, 16-19) The inability to correlate immunogenicity with a specific quality attribute bears the uncertainty of such correlations. Several studies have described methods for fractionation (20-24) that were subsequently used for evaluation of the immunogenicity of various aggregate/particle fractions in in vitro or in vivo model systems. However, despite the valuable insights that such studies have provided with regard to the impact of particle size, the potential influence of the presence of chemical modifications has remained largely unaddressed and overlooked. One exception is a recent investigation published by our team using a human IgG1 transgenic mouse model and thoroughly characterized soluble IgG1 aggregates ranging from dimers to 20-mers. (25) In this study, we presented the generation of well-defined size fractions of subvisible particles of a) unmodified IgG1 and b) extensively oxidized IgG1 by a variety of chemical treatments, using the 3 ACS Paragon Plus Environment


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 22

recently reported method of fractionation by fluorescence-activated cell sorting (FACS).(24) A previously characterized IgG1 transgenic (tg) mouse model (hIgG1 tg) (25) was used to evaluate unfractionated subvisible particle populations and particle fractions with an approximate diameter of 15 µm for their ability to induce immune responses. All samples used in the study were characterized thoroughly with regard to their size distribution, biophysical properties and chemical modifications.


Page 5 of 22

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


Material and Methods Two IgG1 monoclonal antibodies (mAb1 and mAb2) were selected as model proteins for this study. The mAbs, both IgG1 type , were provided by F. Hoffmann-La Roche Ltd. (Basel, Switzerland). The protein concentration of mAb1 in solution was 25 mg/ml; mAb2 (used as confirmatory model) was 100 mg/mL. Prior to further handling, the clinical grade drug products were filtered through a 0.22 mm Millex GV (PVDF) syringe filter units (Millipore, Bedford, MA). All other chemicals used were of analytical grade and obtained from commercial sources. Stress Conditions Thermal stress was applied with a thermomixer (Eppendorf, Germany) as described previously (24). For mAb1, 1 ml of the 25 mg/ml solution was incubated for 3 min at 80 °C with 1400 rpm shaking. MAb2 was diluted to 25 mg/ml with phosphate buffered saline (PBS) before incubation at 90 °C for 10 min. Stressed solutions were then resuspended by 20 consecutive successive aspirations and expulsions using a disposable Norm-inject 5 ml luer lock silicone free syringe (Henke Sass Wolf, Tuttlingen, Germany) with an attached 27 Gauge needle (0.4 x 40 mm) (Braun, Melsungen, Germany). The stressed unfractionated solutions were stored at -80 °C. For generating chemically modified samples, the protein solutions were treated with UV-VIS light or chemicals before applying heat or stir stress. For UV stress, glass vials containing 16 ml formulated antibody solution were kept under the UV–VIS light (irradiance energy of 765 W/m2) for 30 h using a SunTest XLS+ device (Atlas-MTS, Mount Prospect, Il). Chemical stress was applied by adding 2,2azobis(2-amidinopropane) dihydrochloride (AAPH) or H2O2 to the mAb solution. Samples were incubated under protection from direct light with either 5 % AAPH at 40 °C for 120 h (hereafter referred as AAPH1 samples) or 1 % H2O2 at 5 °C for 24 h. Incubation with AAPH in combination with free methionine resulted in selective tryptophan oxidation as described in detail by Folzer et

al.(26) (hereafter referred as AAPH2 samples). Preparation of size fractions of mAb particles by Preparative Flow Cytometry (FACS) For the preparation of the subvisible particle fractions, a BD FACS Aria III preparative cell sorter (BD Biosciences, San Jose, California) was used with BD FACSDiva v 6.1 software as described previously (24). Additionally, a highly pure sorting modality (four-way purity sorting for FACS Aria) was applied, achieving a purity of 98 % with a yield >80 % for event rates used in this study. Subvisible particle characterization Light obscuration (LO), Flow Imaging microscopy (FI) and Resonant Mass Measurement (RMM) were performed as described previously (24).



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

Page 6 of 22

Peptide Map Analysis and proteolytic digest of mAb For detection and quantification of oxidized amino acids, the mAbs were digested with trypsin. The tryptic peptide mixture was afterwards separated by reversed phase-UPLC. This methodology has been previously reported. For additional details, please refer to (26). Additionally, peptides of interest were identified based on their expected m/z-values. For their quantification, GRAMS AI software was used (Thermo Fisher Scientific). Relative amounts of non-oxidized and oxidized peptides were calculated by manual integration of the corresponding peaks. Capillary Electrophoresis Sodium Dodecylsulfate-NonGel Sieving (CE-SDS-NGS) CE-SDS-NGS was performed both under reducing, and non-reducing conditions. Samples were treated with SDS (and dithiothreitol for reducing conditions) for 10 min at 70 °C before analysis by capillary electrophoresis (ProteomeLab PA800 Beckman Coulter, Nyon, Switzerland). The capillary was rinsed at 70 psi with 0.1 mM NaOH for 5 min and then with 0.1 mM HCl and deionized water for 1 min. The SDSMW Gel Buffer was loaded at 50 psi for 15 min. All samples were labelled with 3-(2furoyl)quinoline-2-carboxaldehyde (Molecular Probes). Samples were injected electrokinetically at 5 kV for 30-40 s. Analysis was performed in the negative polarity mode (−15 kV, −480 V/cm). Endotoxin Endotoxin levels were measured in all samples using a Limulus Amoebocyte Lysate (LAL) assay kit Limulus Amebocyte Lysate (LAL) Kinetic-QCLTM, (LONZA, Walkersville, MD) according to the manufacturer’s instructions and method from PhEur (kinetic chromogenic), chapter 2.6.14. FTIR Infrared spectra were acquired using a Nicolet 6700 Fourier transform infrared spectrometer (Thermo Electron Corporation, Beverly, MA). Samples were placed on a 0.8 µm size pore filter (gold coated polycarbonate Rap.ID Particle Systems GmbH, Berlin Germany) connected to vacuum and washed with particle-free water. Spectra were obtained used the ATR (attenuated total reflection) mode of the FTIR microscope. A 64-scan interferogram was collected for each spectrum in single beam mode with a 4 cm−1 resolution. Protein concentration was approx. 25 mg/ml for all the samples subjected to FTIR analysis. Data were analyzed with the OMNICTM software version 8.3 tool and displayed as second derivative amide I spectra with a 25-point smoothing. TEM


Page 7 of 22

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


For TEM experiments, 4 microliters of the particle solutions were placed onto Lacey Carbon 300 Mesh Copper grids (Ted Pella, Redding, CA) and allowed to adsorb for 60 s. Grids were negatively stained using 2 % uranyl acetate (pH 4.3) solution. Dried grids were examined by TEM. Samples were analyzed at a nominal magnification of 3400x using a Philips CM10 electron microscope (Philips, Eindhoven, The Netherlands) operating at 80 kV. Mice and Immunization The transgenic mouse model (hIgG1 tg ) was validated recently, as reported by Bessa et. al. (25) The human IgG tg mice (C57BL/6 background) were hemizygous for the human IgG construct. Wild type C57BL/6 littermates were used as controls. Experiments were conducted in accordance with the Swiss legislation on animal welfare and were approved by the local authorities. IgG1 particle samples were administered subcutaneously twice per week. A total of 140 µg purified particles (10 µg per injection) were administered at the level of the abdomen. ELISA Anti-drug antibodies (ADAs) were measured using ELISA plates (Nunc Immuno MaxiSorp, Rochester, NY)coated overnight with the respective native mAb (5 µg/ml) that was used to generate oxidized material and particles. ELISA was performed according to standard protocols (25).



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 22

Results Generation and biophysical characterization of subvisible proteinaceous particle fractions The aim of this study was to evaluate the immunogenicity of subvisible proteinaceous particles in a hIgG1 transgenic mouse model. Fractions of proteinaceous particles with 15 µm Equivalent Circular Diameter (ECD) were created and compared with the unfractionated sample containing proteinaceous particles with diameters ranging from hundreds of nanometers to approximately one hundred micrometers. In addition, different modifications in the protein primary structure were generated by UV irradiation or addition of chemical oxidants (26). Unfractionated as well as enriched 15 µm particles obtained using a recently-described FACS method were tested (24). Particles were sorted applying gate limits defined by certified bead size-standards. Light obscuration (LO), Flow Imaging (FI), flow cytometry and Resonant Mass Measurements (RMM) were used to measure the particle size distributions of the resulting fractions. Table I summarizes the characteristics of these fractions and the unsorted starting material that was used to create them. Figure 1 shows the particle sizedistributions of all samples used in this study measured by FI in the range of 1 to 100 µm and RMM in the range of 0.013 to 5 µm. Representative FI images are shown in Supplementary Figure 3. The mass distributions of all particle samples were calculated using digital image morphology analysis (for measurements using Flow imaging microscopy) (27) or the buoyant mass (RMM measurements) (28) (Figure 1). The calculated mass distributions demonstrated that most fractions contained some contaminating particles in the nanometer size range that co-purified during FACS separation. However, the impact of such carryover was negligible in terms of protein mass. Thus, the size distribution of the fractions showed the highest particle count around 1 µm, but the mass distribution was shifted to larger sizes of 15 µm (Figure 1). This can be explained by the power relationship ସ ଷ

between size and volume (for a sphere = ߨ‫ ݎ‬ଷ ), i.e. a small number of particles of larger size contain significant amounts of protein. This observation was even more pronounced for the unfractionated material that contained millions of particles smaller than 1 µm, but also particles up to 60 µm, which comprise most of the protein mass. As expected for the unfractionated samples, the particles were broadly distributed over size both in terms of count and mass. Additional biophysical characterization of all samples revealed no discernable differences either in terms of ultra-morphology (measured by transmission electron microscopy – TEM) or higher order structure (measured by Fourier transform infrared spectroscopy – FTIR). TEM analysis showed that all particles exhibited a similar morphology. There were no discernible differences between differently stressed materials (e.g. different types of oxidation stress). Supplementary Figure 2 contains representative TEM images at a resolution of 500 nm with protein particles represented by the black to dark-gray regions. Notably, TEM analyses confirmed the size-distributions measured using FI or RMM. The FTIR spectra of all samples were highly similar as well (see Supp. Fig 2 for 8 ACS Paragon Plus Environment

Page 9 of 22

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


representative spectra), indicating that no major structural differences between the different fractions were detected. Characterization of chemical modifications Chemical modifications were analyzed by tryptic peptide mapping using liquid-chromatography-mass spectrometry (LC-MS) or capillary sodium dodecylsulfate electrophoresis (CE-SDS) under reducing and non-reducing conditions. The LC-MS analysis (Figure 2) showed that the unoxidized fraction contained only very low levels of chemical modifications. This was similar to the unfractionated (and unoxidized) starting material. The sorting process thus had a negligible influence on the protein primary structure. Not surprisingly, increasing harshness of the treatment by UV, H2O2 and AAPH, increased the relative oxidation of methionine and tryptophan residues in the protein. AAPH treatment led to the highest tryptophan oxidation levels, whereas H2O2 treatment resulted mostly in methionine oxidation. Protection of methionine residues by addition of free methionine (sample AAPH2) resulted in a significant reduction of oxidation for several methionine residues, i.e. M258, M364 and M434, in the heavy chain. In the UV-treated sample, methionine and tryptophan oxidation were also be observed, but the relative levels were lower than in the chemically oxidized samples. It needs to be pointed out that in the peptide mapping analyses reported here, only common modifications (such as Trp and Met oxidation) can be readily detected. (26) Obviously, oxidative degradation of proteins results in a large variety of reaction products. Indeed, the peptide map chromatograms showed a number of unidentified peaks, increasing in the more strongly oxidized samples (e.g. AAPH1). However, besides the summarized large number of methionine and tryptophan oxidation sites (Figure 2; Supplemental Figure 2), no additional oxidative modifications (such as histidine oxidation or diTyrosine formation) were detected, despite careful interrogation of the experimental mass spectra. The increasing levels of oxidation (i.e. covalent modifications) could also be followed by CE-SDS analysis. Figure 3 shows the electropherograms of unfractionated and fractionated samples under nonreducing (3A) and reducing (3B) conditions. The unfractionated and the unoxidized 15 µm fraction showed a similar profile to the starting material. In contrast, the UV treated sample showed a significantly increased high molecular weight peak under both reducing and non-reducing conditions, indicating the formation of intermolecular covalent linkage. The samples that underwent harsher chemical oxidation demonstrated even greater changes in electrophoretic profiles (Figure 3). Interestingly, CE-SDS analyses also revealed significant differences between different chemical treatments. For example, the observed fragmentation of the antibodies was relatively minor in H2O2– oxidized samples and severe in AAPH-oxidized samples. Immune Responses in Human IgG Transgenic Mice



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

Page 10 of 22

After biophysical characterization, the particle samples were used to immunize human IgG1 transgenic (tg) mice. The anti-drug antibodies (ADA) titers after 28 days, as determined by ELISA, are reported in Figure 4. The comparison of samples with different extents of chemical (covalent) modification showed a direct correlation between the level of chemical modification and immunogenicity. Neither the unoxidized unfractionated subvisible particles, nor the 15 μm unoxidized sample broke tolerance in the hIgG tg mice. The fraction that received UV treatment broke tolerance in three out of five tg mice tested, whereas the chemically oxidized material (showing higher levels of covalent modifications) broke tolerance in all five tested mice. Interestingly, the oxidized samples induced increasing ADA titers in tg mice in the order: UV

Extensive Chemical Modifications in the Primary Protein Structure of

Recommend Documents