Specific Enzymatic Digestion and Fragment Isolation for the Critical Quality Attribute Assessment of Igg1

ZURICH UNIVERSITY OF APPLIED SCIENCES DEPARTMENT OF LIFE SCIENCES AND FACILITY MANAGEMENT INSTITUTE OF BIOTECHNOLOGY

Specific enzymatic digestion and fragment isolation for the critical quality attribute assessment of IgG1

Master thesis

by Rappo, Martin Alexander Master of Life Sciences 20.10.2017

Pharmaceutical Biotechnology

Academic supervisor: Prof. Dr. Jack Rohrer Head of the Cell Physiology and Cellular Engineering Group ZHAW Life Sciences und Facility Management, Einsiedlerstrasse 31, 8820 Wädenswil, Switzerland

Industrial supervisor: Dr. Christoph Rösli Senior Fellow in the Physicochemical Analytics Group Novartis Pharma AG, Werk Klybeck, Postfach, 4002 Basel, Switzerland

Imprint

Keywords

Critical Quality Attributes, CQA, Monoclonal Antibody, mab, Fab, Fc, F(ab’)2, Large hinge IgG fragment, LHF, IdeS, IgdE, Lysine Gingipain, Kgp, EndoS2

Please cite this report as followed:

M. A. Rappo, Specific enzymatic digestion and fragment isolation for the critical quality attribute assessment of IgG1, ZHAW LSFM (2017)

Contact information

Author

Martin Alexander Rappo Speiserstrasse 12 4600 Olten Switzerland Tel. +41 79 537 64 76 E-Mail: [email protected]

Academic Supervisor

Prof. Dr. Jack Rohrer ZHAW Life Sciences und Facility Management Einsiedlerstrasse 31 8820 Wädenswil Switzerland Tel. +41 58 934 57 17 E-mail: [email protected]

Industrial Supervisor

Dr. Christoph Rösli Novartis Pharma AG Werk Klybeck Postfach 4002 Basel Switzerland Tel. +41 61 696 14 39 E-Mail: [email protected]

Basel 20.10.2017

iii

Acknowledgement

I would like to express my very great appreciation to Christoph Rösli. I would like to thank him for giving me the chance to work on this interesting project, the stimulating scientific discussions and useful inputs given during the time I spend at Novartis.

I am particular grateful for the support and training given by Olivia Guerre. Also the interesting exchange helped to approach problems with a different perspective.

I would like to offer my special thanks to Philippe Simeoni, who always offered this time whenever a question or issue regarding CE came up.

Also Francois Griaud I would like to thank for giving me an introduction on antibody digestion.

A thanks also goes to Wolfram Kern for being so kind to let me use his work bench and all his equipment.

The refreshing exchange with Manuel Diez inside and outside the office has been a great help to keep motivated and stay open-minded.

I would like to thank Genovis for kindly providing me a free test sample of FabALACTICA™.

A special thanks also goes to Raphael Dreier for all the coffee and lunch breaks we had together, the support during stressful times, extended brain-storming sessions and all the laughs we shared.

Abstract

Recent ICH quality guideline updates introduced the enhanced drug development which is based on Quality by Design approach. For this a Quality Risk Management has to be conducted based on scientific knowledge. Information gaps are closed through analysis of molecules with domain-specific modifications. The aim of this thesis was to evaluate methods to prepare a model IgG1 for such studies. The antibody fragments F(ab’)2, LHF, Fab and Fc and the Fc-glycosylation cleaved IgG with residual N-acetylglucosamine were selected. A representative IgG1 was digested with the recently discovered enzymes IdeS, Kgp, IgdE and

EndoS2 and the digestion products were purified with common biotechnological down-stream methods. The F(ab’)2 was successfully isolated with a yield of 55% and a purity of 96%. After an initial digestion parameter optimization the LHF, Fab and Fc could be generated and an isolation method was established. A following scale-up was unsuccessful, due to reproducibility issues. A promising alternative was identified. The enzymatic glycan cleavage and product isolation was successful with a yield of 80% and a purity of 99%. Zusammenfassung

Die kürzlich von der ICH aktualisierten Qualitätsrichtlinien führten den erweiterten Wirkstoffentwicklungsvorgangs ein, welcher auf dem Quality-by-Design Prinzip basiert. Dazu wird ein Risikomanagement, basierend wissenschaftlichen Fakten durchgeführt. Um Informationslücken zu schliessen müssen proteindomänenspezifisch veränderte Moleküle untersucht werden. Das Ziel dieser Arbeit war die Evaluation von Methoden zur Vorbereitung eines Modell IgG1 für solche Studien mit minimaler unspezifischer Veränderungen.

Ausgewählt wurden die IgG-Fragmente F(ab‘)2, LHF, Fab und Fc und das Fc-Polysaccharid gespaltene IgG mit restlichem N-Acetylglucosamin. Ein repräsentatives IgG1 wurde mit den kürzlich entdeckten Enzymen IdeS, Kgp, IgdE und EndoS2 verdaut und mittels üblichen biotechnologisch Reinigungsmethoden aufgereinigt. Das F(ab‘)2 Fragment konnte mit einer Ausbeute von 55% und Reinheit von 96% isoliert werden. Nach einer initialen Optimierung der Verdauungsparameter konnte das LHF, Fab und Fc erfolgreich hergestellt und eine Methode zur Fragmentisolation gefunden werden. Die Herstellung einer grösseren Menge war erfolglos, jedoch konnte eine alternative Lösung identifiziert werden. Die enzymatische Spaltung des Polysaccharids und Isolation war erfolgreich mit einer Ausbeute von 80% und einer Reinheit von 99%.

Table of content

1 INTRODUCTION ...... 1 2 THEORETICAL BACKGROUND...... 5 2.1 ANTIBODIES ...... 5 2.1.1 Structure ...... 5 2.1.2 Fragment nomenclature ...... 7 2.1.3 Glycosylation ...... 8 2.1.4 Mechanism of action ...... 8 2.2 ENZYMES ...... 8 2.2.1 Immunoglobulin-degrading enzyme of Streptococcus pyogenes ...... 9 2.2.2 Lysine gingipain of Porphyromonas gingivalis ...... 11 2.2.3 Immunoglobulin G degrading enzyme of Streptococcus agalactiae ...... 13 2.2.4 Endo-beta-N-acetylglucosaminidase of Streptococcus pyogenes ...... 13 2.3 KINETICS MODEL FOR THE KGP DIGESTION ...... 15 3 METHODS AND MATERIALS ...... 17 3.1 INSTRUMENTS, CONSUMABLES, REAGENTS AND SOFTWARE...... 17 3.2 MODEL ANTIBODY ...... 17 3.3 BUFFER PREPARATION ...... 17 3.4 PHYSICOCHEMICAL METHODS ...... 18 3.4.1 Cation exchange chromatography ...... 18 3.4.2 Size exclusion chromatography ...... 19 3.4.3 Capillary electrophoresis – sodium dodecyl sulfate ...... 20 3.5 DIGESTION AND ISOLATION ...... 22 3.5.1 Fab, Fc and LHF ...... 22 3.5.2 F(ab’)2 ...... 27 3.5.3 Chitobiose cleavage ...... 29 3.6 ECOLOGY, SAFETY AND DISPOSAL ...... 30 4 RESULTS AND DISCUSSION ...... 31 4.1.1 Fab, Fc and LHF ...... 31 4.1.2 F(ab’)2 ...... 44 4.1.3 Chitobiose cleavage ...... 46 5 CONCLUSION ...... 48 6 OUTLOOK ...... 49 7 REFERENCES ...... 50 7.1 LIST OF ILLUSTRATIONS ...... 55 7.2 LIST OF TABLES ...... 56 7.3 LIST OF EQUATIONS ...... 57 ANNEX ...... I

Abbreviations

Table 1: Abbreviations.

Acronym Description Ab Antibody ADCC Antibody-dependent cellular cytotoxicity CDR Complementary determining region CE Capillary electrophoresis

CH1 First constant heavy chain domain

CH2 Second constant heavy chain domain

CH3 Third constant heavy chain domain

CL Constant light chain domain DB Digestion buffer DP Drug product Fab Fragment antigen binding Fc Fragment crystallizable FcRn Neonatal Fc receptor FcγR IgG-Fc receptor Fuc Fucose Gal Galactose GAS Group A Streptococcus GlcNAc N-acetylglucosamine HC Heavy chain HPLC High performance liquid chromatography ICH International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use Ig Immunoglobulin IgdE Immunoglobulin G degrading enzyme of S. agalactiae LC Light chain Mab Monoclonal antibody Man Mannose NeuGc N-glycosylneuraminic acid NR Non-reducing PAGE Polyacrylamide gel electrophoresis PD Pharmacodynamics

vii

PK Pharmacokinetics QbD Quality by Design QRM Quality risk management Rgp Arginine gingipain RMT Relative migration time Rt Room temperature SDS Sodium dodecyl sulfate

VH Variable heavy chain domain

VL Variable light chain domain

Amino acids Amino acid A Ala Alanine C Cys Cysteine D Asp Aspartic acid E Glu Glutamic acid F Phe Phenylalanine G Gly Glycine H His Histidine I Ile Isoleucine K Lys Lysine L Leu Leucine M Met Methionine N Asn Asparagine P Pro Proline Q Gln Glutamine R Arg Arginine S Ser Serine T Thr Threonine V Val Valine W Trp Tryptophan Y Tyr Tyrosine

Figure 1: Amino acid properties Venn diagram [1].

viii ZHAW LSFM, 2017 Introduction Martin Alexander Rappo

1 Introduction

With the creation of the first monoclonal antibody (mab) in 1975 a new era of pharmaceutical development was initiated [2]. The mab technology increased therapeutic options and opened new treatment opportunities for diseases with an unmet medical need, such as cancer, immunological disorders, chronic inflammatory disease, infectious disease and many more [3]. Because mabs have a higher target specificity and reduced systemic toxicity compared to prior predominant small molecular drugs, focus of drug research and development slowly shifted toward this technology [4]. In 1986, the first mab was licensed [2,5]. Sixteen years later, in 2002 the first human mab was approved by the U.S. Food and Drug Administration for pharmaceutical use in humans [3]. In 2014 the annual global mab market was approximately $20 billion, generated by roughly 30 products [2] and is estimated to grow to $125 billion with around 70 products on the market by the year 2020 [5]. In parallel with the transition from small to large molecular drug research and development, regulatory authorities introduced the new quality guidelines Q8 – Q12 for drug development, listed in Table 2 [6–8].

Table 2: List of new ICH quality guidelines.

ICH Guideline Topic Q8(R2) Pharmaceutical development Q9 Quality Risk Management Q10 Pharmaceutical Quality System Q11 Development and Manufacture of Drug Substance Q12 Technical and Regulatory Considerations for Pharmaceutical Product Lifecycle Management

The new guidelines propose an enhanced approach for drug development compared to the traditional. Where the traditional way defines set values and operational ranges for process parameters, based on tests and reproducibility studies to reach their acceptance criteria the enhanced approach integrates quality directly into the process [8]. This is achieved by elaborating a quality risk management (QRM) based on scientific evaluations throughout the whole product lifecycle, a so called quality by design (QbD) approach. The QbD, illustrated in Figure 2, includes a summary of characteristics of a drug product (DP) for its ideal quality, the so called quality target product profile (QTPP). The QTPP contains for example information regarding the indication, the mechanism of action, the dosage form or the dosage strength [9]. Based on the QTPP a critical quality attribute (CQA) assessment is performed, listing any physical, chemical, biological, or microbiological property of a DP having an influence on the quality with regard to activity/potency, pharmacokinetics/-dynamics (PK/PD), immunogenicity and/or safety. Examples of CQAs are viral impurity, host cell proteins, glycosylation or high molecular weight species. Afterwards operation parameters affecting CQAs, the critical

1 ZHAW LSFM, 2017 Introduction Martin Alexander Rappo process parameters (CPP), are defined and tested by a design of experiment (DoE). Based on the results of these experiments, the criticality of the CQAs is adapted. Examples for CPPs are: temperature, time and pH. Based on the findings a working area of each CPP is defined, called the design space. At last a control strategy is introduced, ensuring the defined quality of the final DP. The whole process is maintained and updated through the whole lifecycle and should lead to an improved product understanding and ultimately higher patient safety [10– 14].

Quality target product profiles (QTPP)

Critical quality attributes (CQA)

Critical process parameters (CPP) Design Verification Space

Iterative process throughout the lifecycle Control Validation Strategy

Figure 2: Illustration of the Quality by Design strategy [15,16].

To conduct a QRM potential risks are identified first, their likelihood is rated and the consequences are assessed. Based on those points an impact factor for each risk is calculated and an appropriated follow-up action is evaluated. Such can be “no action required”, “further investigation required”, “close monitoring” or “risk prevention”. The basis for such a risk identification for quality attributes emerges from four sources, illustrated in Figure 3: literature & in-silico data, in-house knowledge and product comparison, project related experimental data and specific clinical experience. The information availability is usually high for literature & in-silico data and decreases for in-house knowledge & product comparison. Project related experimental data and specific clinical expertise, which have both a high certainty score and cost & time investment, are often only available in late development phases. Therefore, the relation of effort to information gain has to be evaluated thoroughly.

2 ZHAW LSFM, 2017 Introduction Martin Alexander Rappo

Literature & In-house knowledge Project related Specific in-silico data & product comparison experimental clinical data experience

Certainty Availability

Cost and duration

Figure 3: Potential information sources for the quality risk management [15].

The risks considered for impact assessment of quality attributes during drug development are divided into four categories, listed in Figure 4.

Biological activity/potency describes the ability of a molecule/drug to induce a certain response in a biological system at a defined concentration [17].

Pharmacokinetics (PK) is the study of progression of a drug during absorption, distribution, metabolism and excretion. Pharmacodynamics (PD) is the study of the biological effect caused by the drug at the site of action [18].

Immunogenicity describes the ability of a substance inducing an immunological reaction in an organism This includes the formation of anti-drug antibodies [19].

Safety according to the World Health Organization are biological, chemical or physical agents or operations that are reasonably likely to cause illness or injury [20].

Figure 4: The four quality categories for the risk assessment quality attributes [15,17–20].

This QbD approach slowly finds its way into pharma companies, where chemistry, manufacturing and control strategies are adapted and updated accordingly. For an adequate risk evaluation a fundamental molecule, mechanism and process understanding is necessary. During the CQA assessment related to the active pharmaceutical ingredient in-depth understanding of potential structural and molecular modifications and their impact are evaluated. As literature and in-silico data are generally rated quiet uncertain and are often not

3 ZHAW LSFM, 2017 Introduction Martin Alexander Rappo even available for new biological entities, molecular entity-specific in-house knowledge has to be expanded and tools to gain project-specific experimental data have to be established. So far biopharmaceuticals have been exposed to controlled stressed conditions, like heat, oxidation or shearing stress, for a defined period of time and degradation products were analyzed. This approach may introduce multiple, uncontrolled modifications leading to vague conclusions. Therefore, the development of new tools to modify a drug very specific (and prevent undesired modifications) is necessary.

The topic of this study is to start establishing tools for the CQA assessment of biopharmaceuticals. Potential modification are for the most part dependent on the molecular structure of the drug substance and may vary from one project to another. As a starting point the immunoglobulin G (IgG) 1 subclass was used in this thesis. IgG1 is currently the most used subclass of mabs in drug development. A possibility of targeted modification is obtained through enzymes. Especially in protein-related sciences enzymes offer a wide variety of applications. Unfortunately, commonly used endoproteinase, such as trypsin, Lys-C, papain or pepsin still tend to undesired over-digestion. In recent years new endoprotease were discovered which promise to cleave IgGs sequence specific without the risk of over-digestion [21–24]. The goal of this thesis is to screen and improve the enzymatic digestion of a model

IgG1 to obtain Fab, Fc, LHF and F(ab’)2 fragments, respectively the chitobiose cleaved IgG, shown in Figure 5, and isolated the corresponding products. Undesired modifications have to be prevented as far as possible. For a copy of the objective setting agreement please see appendix page X.

F(ab‘)2 LHF  Bivalent binding (potency)  Monovalent binding (potency)

Fab Fc  Monovalent binding (potency)  Glycosylation pairing (PK/PD)  Fc methionine oxidation distribution

(Fuc-)GlcNAc-IgG

 Change in higher order structure  FcRn, FcγR and C1q binding (PK/PD)

Figure 5: Selected digestion products to be evaluated during the master thesis including their final purpose in the CQA assessment.

4 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

2 Theoretical background

2.1 Antibodies

Antibodies (Ab), also called immunoglobulin (Ig), are glycoproteins and play an essential role in the adaptive immune system. Abs are naturally either transmembrane proteins, as B-cell receptors or soluble proteins released from plasma cells after an antigen recognition. Special properties of such Abs are the high specificity and affinity to bind a certain antigen and initiate an effector mechanism [25].

2.1.1 Structure

The basic structure of an Ab, illustrated in Figure 6, is Y-like shaped and consists of the stem, the so called fragment crystallizable (Fc) responsible for the effector function (explained in chapter 2.1.4) and the two arms, the so called fragment of antigen-binding (Fab) responsible for the specific antigen binding [26]. Abs are axial symmetrical hetero-tetramers containing two light chains (LC) and two heavy chains (HC). The LC consist of two domains of roughly

12.5 kDa, the variable (VL) and constant (CL) domain. The HC consists of four domains, also one variable (VH) and three constant domains (CH1, CH2 and CH3). The two HC are interlinked between CH1 and CH2, in the so called hinge-region via disulfide bridges [25,27]. Similarly, the LC and the HC are liked via a disulfide bridge in the Fab region [25]. Additionally, each domain contains an intra-chain disulfide bond [28]. The CH2 bears an N-glycosylation at position Asn297, which plays an important role for the effector function, protein stability and the ability to aggregate [29,30]. The highly specific binding ability of Igs origins from the hypervariable regions within the variable region, the so called complement determining region (CDR) [25]. Each variable domain contains three CDRs and the CDR of the HC and LC lie close together through the three-dimensional arrangement forming the so called antigen-binding pocket [31].

5 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

IgG1 structure Heavy chain Light chain

Variable region Fragment antigen-binding (Fab): antigen binding

Hinge region: flexibility and connection

Constant region Glycosylation Fragment crystallizable

(Fc): Effector function

Figure 6: Illustration of the IgG1 structure with light chain in apricot and heavy chain in blue.

Fab region with hapene in IgG1 3D structure binding pocket

LC HC

Glycosylation

Figure 7: Ig Fab with CDR in red and a hapten as Figure 8: 3D surface structure of an IgG1 [33]. epitope [32].

6 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

In humans there are two different types of LCs, the κ and the λ chain and five different types of HCs, the γ, δ, ε, α and μ chain. Depending on the HC type, antibodies are classified into five classes or isotypes, namely IgG, IgD, IgE, IgA and IgM [25–27]. IgG have the highest half-life of roughly three weeks and are therefore the most interesting class for therapeutic applications [25]. The IgG class is further divided into four subclasses [25,27,34,35], depending on the preferable antigen target, hinge region length, disulfide connection [26,28], serum concentration, half-life, placental transfer potential, complement activation and Fc receptor binding. Those subclasses are numbered ascending from highest to lowest relative abundance (γ1, γ2, γ3, γ4) in human serum. IgG1 is usually found in the highest relative abundance in the human serum compared to other subclasses and targets mainly soluble and membrane proteins. Additionally, it has compared to other subclasses a unique disulfide connection between LC and HC [34] and binds efficiently to the IgG-Fc receptor (FcγR) [36].

2.1.2 Fragment nomenclature

In Figure 9 common antibody fragments are listed.

Immunoglobulin G Large hinge fragment [40] F(ab‘)2 IgG LHF

Reduced/half IgG Fab‘ Fragment antigen binding rIgG/hIgG Fab

Fd‘ Fragment crystallizable Fc/2 Fc

The term Fabc is rarely used and can have two different meanings. Either as equivalent for the intact IgG [37,42] or the IgG without CH3 domains [41].

Figure 9: Nomenclature of common IgG fragments [37–42].

7 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

2.1.3 Glycosylation

Protein glycosylation is a highly complex post-translational modification but crucial for folding, conformation, stability & half-live, solubility, pharmacokinetics, biological activity and immunogenicity [29,43]. Ig of the subclass γ1 undergo glycosylation in the endoplasmic reticulum and Golgi network and contain a conserved glycosylation at approximate position

297 Asn of both CH2 domains in the Fc region. Roughly 20% of mabs contain an additional N- glycosylation in the variable region. Usually glycosylation of Abs consists of a complex bi- antennary glycan with core-fucosylation [44]. An illustration of transgenic animal respectively Chinese hamster ovary cell N-glycosylation structures is shown in Figure 10. Minor size and structural differences of the glycan from one protein to another lead to an unequal mixture, which is called glycoform heterogeneity [44,45]. Specifically for IgG the glycosylation has an effect on thermo-, unfolding- and protease-stability, impacts the aggregation tendency [46] and influences the effector functions, especially the antibody-dependent cell cell-mediated cytotoxicity (ADCC) and the complement dependent cytotoxicity (CDC) [29,46,47].

Fc glycosylation of Chinese hamster ovary cells

Chitobiose core

Figure 10: Fc glycosylation structure of transgenic animal/Chinese hamster ovary cells.

2.1.4 Mechanism of action

Abs can have different effects on an antigen. Important for a functional Ab is the ability to bind a certain epitope with the binding pocket and initiate an effector function from the Fc region. With both of these properties Abs act as adapter and bring antigen and the immune cell together. Especially IgG1 and IgG3 are known to initiate such effector mechanisms [34]. Therapeutic mabs on the other hand usually do not depend on an effector function and in some cases those are even undesired. A summary of the most important effector functions are listed in the appendix on p. II.

2.2 Enzymes

Enzymes are proteins with a wide function of catalyzing bio-chemical reactions in an organism. Some enzymes catalyze a certain reactions high specific and under physiological conditions.

8 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

Enzymes also have a high relevance in industrial and diagnostic applications. A summary of commonly used enzymes for the mab analytic is listed in the appendix on p. III.

2.2.1 Immunoglobulin-degrading enzyme of Streptococcus pyogenes

Name Immunoglobulin-degrading enzyme of Streptococcus pyogenes Catalytic domain Abbreviation IdeS, streptococcal Mac-1 L R Catalytic type Cysteine [48] Family Protease C66 [49]; CA clan [50] Organism Streptococcus pyogenes

Molecular weight 36.2 kDa [50] Figure 11: Illustration of IdeS. Molar absorbance 1.362 (calculated) [51]

Streptococcus pyogenes is a bacteria causing strep throat and impetigo in humans [52]. Recent studies revealed its secreted cysteine endopeptidase (IdeS or streptococcal Mac-1) to naturally protect the bacteria from the host’s immune system by cleaving surrounding Ig at the lower hinge region.

Figure 12: IgG fragmentation of IdeS.

IdeS contains similar structural features as papain but shows only minor sequence homology. Both can structurally be divided into a left-hand (L) and right-hand (R) domain and the catalytic site is located at the interface in between. High specificity of substrate recognition towards the hinge region of IgG is a unique feature of IdeS. So far only few accepted sequences of the Ig hinge region have been identified, shown in Figure 13 [50]. This high specificity, which prevents over-digestion is exploited for bioanalytical applications.

9 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

Figure 13: Accepted sequence for the IdeS cleavage. Identical amino acid in black, closely chemically similar in blue, weakly chemically similar in sienna and non-similar in apricot.

The catalytic mechanism proposed for the IdeS digestion is similar to the one of the papain cysteine protease superfamily, forming a thiolate-imidazolium ion with the catalytic dyad Cys94 and His262, shown in Figure 14 [50].  Through proton transfer from His262 to Cys94 the enzyme is activated.  The thiolate of Cys94 attacks the carbonyl of the substrate and forms an enzyme-substrate complex.  The scissile peptide bond of the substrate is broken and the C- terminal fragment is released, where the remaining chain forms together with the enzyme a thioester, the so called acyl-enzyme intermediate.  The thioester is hydrolyzed in a subsequent step and the N-terminal chain is also released [53].

 

 

Figure 14: Proposed catalytic cycle of the IdeS cleavage.

Asp284 and Asp286, close to the catalytic pocket, have been identified to have a positive impact on the catalytic activity, by facilitating the correct three-dimensional orientation resp. creating

10 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo a suitable electrostatic milieu. Additionally, Lys84, which is stabilized through the hydrogen bond-salt link of Asp286, forms together with the active site Cys94 a stabilizing pocket for the substrate, the so called oxyanion hole [50]. This is illustrated in Figure 15.

Figure 15: Catalytic pocket and stabilizing effects.

Initial kinetic studies showed a non-Michaelis-Menten behavior, but a curve of positive cooperativity [54]. Later studies revealed an underlying two-step Michaelis-Menten kinetics, where the first reaction is 100 times faster than the subsequent [55].

Genovis®, a life science company, offers recombinant IdeS in various forms, e.g. as lyophilisate (FabriCATOR®), immobilized on agarose (FragIT™) or in combination with a CaptureSelect™ column (FragIT™ kit) including digestion procedure.

2.2.2 Lysine gingipain of Porphyromonas gingivalis

Name Lysine gingipain Hole with catalytic pocket Abbreviation Kgp Enzyme class EC 3.4.22.47 [54] Catalytic type Cysteine Family Protease C25; CD clan [53] Organism Porphyromonas gingivalis

Molecular weight 187 kDa (native) [56] Immunoglobulin superfamily- 48 kDa (recombinant; catalytic like domain and immunoglobulin superfamily-like domain) Figure 16: Illustration of recombinant Kgp. [53,57]

Molar absorbance 1.971 (recombinant, calculated) [58]

The periodontal disease is a wide spread bacterial infection of the gums and the bacteria supposedly to be the main cause is Porphyromonas gingivalis. P. gingivalis expresses two

11 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo types of cysteine peptidase gingipain, the lysine (Kgp) and the arginine gingipain (RgpA and RgpB). These two enzymes contribute to a large part to the extracellular proteolytic activity responsible for its high virulence [57]. Structurally Kgp consists of five domains; the signal peptide, a pro-domain, the Lys-specific catalytic subdomain, an immunoglobulin superfamily- like domain, and the C-terminal domain [53,57]. Kgp cleaves multiple substrates after Lys [53] and also plays an important role in evading the host’s immune system by degrading Ig at the upper hinge region, illustrated in Figure 17 [54].

Figure 17: IgG fragmentation of Kgp and IgdE.

For Fc glycosylated IgG1 only one cleavage site, above the hinge region, has been identified, shown in Figure 18. For IgG1 without glycosylation a second cleavage site in the Fc region was observed [22]. Also IgG3 have been reported to be cleaved by Kgp but not as specific as IgG1. For IgG2 and IgG4 no cleavage sites have been identified [54].

Accepted Kgp cleavage sequence of IgG1 230 240 .....|...... |...... KSCDK | THTCPPCP...

Figure 18: Accepted Kgp cleavage sequence of IgG1 [22].

The proposed mechanism is similar to IdeS Cys/His dyad involving Cys477 and His444, but experiments have shown the importance of Asp388 for the catalytic cleavage [57]. Typical for cysteine proteases is an activity increase in a mild reducing environment, like glutathione or cysteine [53]. Kinetic studies revealed that the digestion does not follow the Michaelis-Menten kinetic, but rather a mechanism of positive cooperativity [54], usually meaning the first ligand binding to the enzyme simplifies a subsequent binding [59]. Similar to IdeS, Kgp kinetics could also underlie a consecutive two-step Michaelis-Menten kinetic, but this theory was not tested yet. It is reported that the digestion can be stopped by drastically decrease pH or addition of iodoacetic acid resp. Nα-Tosyl-L-lysine chloromethyl ketone [54]. The mayor advantage of Kgp over commonly used papain or Lys-C is the high specificity without the risk of over-digestion. Genovis® offers a recombinant Kgp in a kit including a cysteine solution as reducing agent, called GingisKHAN®.

12 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

2.2.3 Immunoglobulin G degrading enzyme of Streptococcus agalactiae

Name Immunoglobulin G- degrading Enzyme Abbreviation IgdE Catalytic pocket Catalytic type Cysteine Family Protease C113, CA clan [60] Organism Streptococcus agalactiae Molecular weight 70 kDa (recombinant) Molar absorbance 0.584 (calculated) [61] Figure 19: Illustration of IgdE.

The immunoglobulin G degrading enzymes (IgdE) of the Streptococcus species, known for their virulence in several species, were recently discovered and have been identified to help the bacteria evade the host’s immune system by degrading surrounding Ig. Exceptional for those enzymes is their high specificity towards IgG of the streptococci’s main host and distinct cleavage site. One IgdE, degrading human IgG1, origins from S. agalactiae and cleaves the Ab similar to papain, above the hinge region, shown in Figure 20, but without further cleavage. The involved mechanism is the same as for Kgp and is formed by the catalytic triad Cys302 His333 and Asp348. Several applicable cysteine protease inhibitors like, Z-LVG-CHN2 and iodoacetamide have been reported [36,49,55].

Accepted IgdE cleavage sequence of IgG1 230 240 .....|...... |...... KSCDKT | HTCPPCP...

Figure 20: Accepted IgdEagalactiae cleavage sequence of human IgG1 [36].

2.2.4 Endo-beta-N-acetylglucosaminidase of Streptococcus pyogenes

Name Endo-beta-N- acetylglucosaminidase

Abbreviation EndoS2 Catalytic site Family Glycoside hydrolases 18 Organism Streptococcus pyogenes group A (GAS) (Serotype M49) [62] Molecular weight 95 kDa (native) [63] 92 kDa (recombinant) [24]

Molar absorbance 1.044 (calculated) [64] Figure 21: Illustration of EndoS2.

Streptococcus pyogenes, as explained in chapter 2.2.1, possesses another way to evade the human immune system. The bacteria produces, next to IdeS, the endoglycosidase EndoS.

13 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

This enzyme is hydrolyzing the glycosidic bond between the two acetylglucosamine (GlcNAc) in the chitobiose core structure of IgG impairing an effector function and increasing bacterial survival [62].

Figure 22: Chemical structure of the chitobiose core.

The EndoS is conserved in almost all identified S. pyogenes serotypes with the exception of serotype M49 GAS strain. The serotype M49 GAS expresses instead the enzyme EndoS2, with similar function but a sequence homology of only 37% [63]. Both enzymes cleave complex type glycans. However EndoS2 cleaves in addition also high-mannose and hybrid type glycans [65]. All human IgG subclasses and Abs of several species, such as mouse, rat, monkey, sheep, goat, cow and horse have been identified as substrate. Importantly, the enzyme works only on native proteins, which indicates a protein-protein interaction. Through a sequence alignment and later active-site mutation the catalytic Glu186 and several Trp have been identified to be of importance for the digestion [62]. Next to IgG only one other protein, the α1- acid glycoprotein, has been identified to act as substrate and therefore EndoS and EndoS2 do not show general chitinase activity.

Figure 23: Chitobiose cleavage of EndoS2.

® ® Genovis offers an EndoS (IgGZero™) and an EndoS2 (GlycINATOR ) ready-to-use kit, containing the corresponding enzyme with a poly-his tag as lyophilisate or immobilized on agarose.

14 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

2.3 Kinetics model for the Kgp digestion

To simplify result comparison and reduce enzyme consumption a model was applied for studies on Kgp. This model is based on the reaction kinetics of a consecutive first-order reaction as described below [66]. The kinetic order was not experimentally determined but assumed as a first order reaction.

k1 k2

A I P Figure 24: Proposed consecutive reaction scheme of an IgG and Kgp.

The IgG reacts in a first reaction step with the kinetic constant k1 to the LHF and a Fab and in a subsequent step with k2 to two Fabs and one Fc. k1 is calculated using Equation 1 and kinetic constants for different digestion parameter are compared. The higher the kinetic constant the faster the digestion.

Equation 1: Determination of the kinetic constant k1 [66].

1 [퐴]0 푘1 = ln 푡 [퐴]푡

It is assumed that k1 and k2 are almost the same, as neither the enzyme nor the substrate change by a lot. Only the accessibility of the substrate of the second reaction is presumably higher compared to the first and k1 is therefore slightly smaller than k2.

푘1 ≈ 푘2 푘1 < 푘2 The molar concentration of the intermediate product is calculated according to Equation 2.

Equation 2: Calculation of the intermediate product concentration [I].

푘1 −푘1푡 −푘2푡 [퐼] = (푒 − 푒 )[퐴]0 푘2 − 푘1 The molar concentration of the product is calculated according to Equation 3.

Equation 3: Calculation of the product concentration [P].

−푘2푡 −푘1푡 푘1푒 − 푘2푒 [푃] = (1 + ) [퐴]0 푘2 − 푘1 The time point where the intermediate product is highest is calculated according to Equation 4.

15 ZHAW LSFM, 2017 Theoretical background Martin Alexander Rappo

Equation 4: Calculation of the time point of the maximal intermediate concentration is reached.

1 푘1 푡[퐼]푚푎푥 = ln ( ) 푘1 − 푘2 푘2 Transformed back to purity the digestion progress is expected to behave as shown in Figure 25.

Simulated Kgp digestion progress for a consecutive first order reaction kinetic 100% 90% 80% 70% 60% IgG1 50% LHF

40% Fc Purity Purity in A% 30% Fab 20% 10% 0% Time in min

Figure 25: Simulated Kgp digestion progress based on a consecutive first-order kinetic.

16 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

3 Methods and materials

3.1 Instruments, consumables, reagents and software.

High performance liquid chromatography (HPLC) was performed on instruments from Agilent, the capillary electrophorese (CE) on BECKMAN-COULTER instruments and the content was measured on a Nanodrop™ by Thermo Scientific. All enzyme were purchased from Genovis®. Analytical data evaluation was performed on Chromeleon™ 6.8. A detailed list of all instruments, consumables, reagents and software can be found in the appendix on p. IV - VI.

3.2 Model antibody

During this studies a model IgG1 was used. The used mab is developed by Novartis and blinded in this report for the protection of the company’s intellectual property. It is simply referred as mIgG1 standing for model IgG1. The mIgG1 belongs to the IgG1 subclass with a κ LC and the standard mab solution is formulated at approximately 200 mg/mL in 5 mM L- histidine/L-histidine hydrochloride at pH 5.8.

3.3 Buffer preparation

Table 3: List of used buffer.

Name Composition CaptureSelect™ binding buffer 10 mM sodium phosphate, 150 mM sodium chloride pH 7.4 CaptureSelect™ elution buffer 100 mM glycine pH 3.0 CE-SDS NR gel buffer Beckman Coulter gel buffer, Cat. no. 391163 CE-SDS NR sample buffer 100 mM tris base, 10 g/L SDS, pH 7.00 CE-SDS red gel buffer Beckman Coulter gel buffer, Cat. no. 391163 + water (1:1.2) CE-SDS red sample buffer 50 mM tris base, 5 g/L SDS, pH 8.5 cOmplete™ solution 1 Tablet in 50 mL Kgp DB

EndoS2 digestion buffer 10 mM sodium phosphate, 150 mM NaCl, pH 7.4 Formulation buffer 5 mM L-histidine/L-histidine-HCl pH 5.8 IdeS digestion buffer 10 mM sodium phosphate, 137 mM NaCl, 2.7 mM KCl, pH 6.0 IgdE digestion buffer 150 mM sodium phosphate pH 7.0 Kgp digestion buffer 100 mM tris pH 8.0/7.3/6.0 Protein L binding buffer 20 mM sodium phosphate, 150 mM NaCl, pH 7.2 Protein L elution buffer 100 mM sodium citrate, pH 2.0 – 3.5 Reducing solution 20 mM cysteine

17 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

3.4 Physicochemical methods

Several physicochemical techniques were used for sample analysis and fragment isolation. Detection was always by UV and the digestion progress was evaluated according to purity by capillary electrophoresis (CE).

3.4.1 Cation exchange chromatography

Cation exchange chromatography (CEX) is a liquid chromatography technique and belongs to the ion-exchange chromatography. It is used to separate species with a different net surface charge on an ionic stationary phase. It was used to detect the rate of deamidation, shown in Figure 26, by monitoring purity of the acidic species.

Figure 26: Deamidation of asparagine.

During this study a salt based separation gradient was applied with parameter listed in Table 4.

Table 4: Applied CEX parameter.

Parameter Setting Sample injection concentration 1 mg/mL Sample injection volume 50 μL Sample temperature 5°C Mobile phase A 25 mM sodium phosphate, pH 6.0 Mobile phase B 25 mM sodium phosphate, 250 mM sodium chloride, pH 6.0 Flow rate 1.0 mL/min. Column ProPac™ WCX-10

18 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

Column temperature 25°C Detection 220 nm Gradient Time in min %B 0 10 30.0 37 30.1 100 32.0 100 32.1 10 40.0 10

An example chromatogram is displayed in Figure 27.

Example CEX chromatogram of mIgG1

Acidic variants Main Basic variants Absorbance in mAU

12 13 14 15 16 17 18 19 20 21 22 Time in min

Figure 27: Example cation exchange chromatogram.

3.4.2 Size exclusion chromatography

Size exclusion chromatography (SEC) separates species according to their hydrodynamic radius. This technique was used to separate and collect the native IgG1, LHF and Fab fractions using the parameters listed in Table 5.

Table 5: Applied SEC parameter for fraction collection.

Parameter Setting Sample injection amount 40 – 400 μg Sample injection volume 10 – 100 μL Sample temperature 5°C Mobile phase 150 mM potassium phosphate, pH 6.5 Flow rate 0.4 mL/min.

19 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

Column TSKgel G3000SWXL; 2 – 3 columns in series Column temperature 30°C Detection 210 nm Run time 55 – 85 min

3.4.3 Capillary electrophoresis – sodium dodecyl sulfate

CE is, like the polyacrylamide gel electrophoresis (PAGE), a physicochemical separation technique but instead of a planar gel the separation occurs in a thin capillary [67]. Sodium dodecyl sulfate (SDS) denatures the protein and ensures a similar charge to mass ratio of all proteins, thus migration time increases with higher molecular weight. The technique is superior to SEC in regard to peak resolution of fragments but has the disadvantage that sample fractionation is not possible.

Non-reducing conditions

For non-reducing (NR) CE-SDS, the protein sample was pre-diluted to 6 mg/mL in water, 20 μL of this solution were further diluted in 75 μL sample buffer and 5 μL 250 mM iodoacetamide solution were added for alkylation. Afterwards the protein was denatured at 70°C for 10 min, the solution was cooled down to room temperature (rt), centrifuged and analyzed by CE-SDS with parameters listed in Table 6.

Table 6: List of CE-SDS NR parameter.

Parameter Setting Auto sampler temperature 20°C Injection type Outlet Injection voltage 5 kV Capillary length 30 cm Capillary diameter 50 µm Capillary temperature 25°C Polarity Positive Voltage 15 kV Detection 214 nm

An example CE-SDS NR electropherogram of the Kgp digestion, with assigned peaks is shown in Figure 28.

20 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

Example CE-SDS NR electropherogram of the Kgp digestion Absorbance in mAU

15 30 45 60 75 90 105 120 135 150 165 180 195 MW in kDa

Figure 28: Example CE-SDS NR electropherogram of the Kgp digestion.

Reducing conditions

For reducing (red) CE-SDS the protein sample was pre-diluted to 6 mg/mL in water, 20 μL of this solution were further diluted in 75 μL sample buffer and 5 μL 2-mercaptoethanol were added for reduction. Afterwards the protein was denatured at 70°C for 10 min, the solution was cooled down to rt, centrifuged and analyzed by CE-SDS with the parameters listed in Table 7.

Table 7: List of CE-SDS red parameter.

Parameter Setting Auto sampler temperature 20°C Injection type Inlet Injection voltage 5 kV Capillary length 30 cm Capillary diameter 50 µm Capillary temperature 25°C Polarity Reverse Voltage 14.2 kV Detection 214 nm

An example CE-SDS red electropherogram of the EndoS2 digestion, with assigned peaks is shown in Figure 29.

21 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

Example CE-SDS red. electropherogram of the EndoS2 digestion

Undigested Digested Absorbance in mAU

15 17 19 21 23 25 27 29 31 33 35 Time in min

Figure 29: Example CE-SDS red electropherogram of the EndoS2 digestion. 3.5 Digestion and isolation

3.5.1 Fab, Fc and LHF

Dig. Digestion IPC In-process control

AC Protein L affinity chromatography

Buffer exchange BE

IPC In-process control

SEC Size exclusion chromatography

Buffer exchange BE

IPC In-process control

Control Purity analysis

Figure 30: Flow chart of the Fab, Fc and LHF generation and isolation.

Stability evaluation in digestion buffer with and without reducing solution

Increased pH and temperature are usually used to enforce asparagine deamidation in proteins [68] and reducing agents lead to breaking of disulfide bridges and loss of

tertiary structure. As all three conditions are present in the Kgp digestion, specified in the Genovis protocol the rate of deamidation and reduction was evaluated.

22 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

Deamidation stability: 20.0 µL mIgG1 solution were diluted in 788 µL digestion buffer (DB) to 5.00 mg/mL (final pH 7.87), split in two equal portions, kept at rt resp. at 37°C/300 rpm for 1, 2 and 24 h and were analyzed by CEX and CE-SDS NR.

Reduction stability: 20.0 µL mIgG1 solution were pre-diluted in 60.0 µL DB to 50.5 mg/mL and 24.8 µL of this pre-dilution solution were further diluted in 50 µL Genovis reducing solution (RS) and 175.2 µL DB, kept at 37°C/300 rpm for 2 h and analyzed by CE-SDS NR.

Digestion progress of Kgp and influence factors

Enzyme ratio: To assess the initial digestion rate and test the applicability of the model, the initial digestion was performed according to the provided protocol from

Genovis [22]. 20.0 µL mIgG1 solution were diluted in 60.0 µL DB and 24.8 µL (1’250 µg) of this solution were further diluted in 50.0 µL DB, 125.8 µL Kgp 10 U/µL (1’258 U) and 50.0 µL fresh RS to 5.00 mg/mL mIgG1. The digestion solution was incubated at 37°C/300 rpm for 5, 15, 30, 60 and 120 min and tested by CE-SDS NR. The 5 min sample was injected twice to prove that digestion stops in presence of 1% SDS.

In the next experiment, the enzyme amount was reduced by a factor of ten. 20.0 µL mIgG1 were diluted in 60.0 µL DB and 24.8 µL (1’250 µg) of this stock solution were further diluted in 162.7 µL DB, 12.5 µL Kgp 10 U/µL (125 U) and 50.0 µL fresh reducing agent to 5.00 mg/mL. The digestion solution was incubated at 37°C/300 rpm for 5, 15, 30, 60 and 120 min and tested by CE-SDS NR. After first data evaluation the experiment was repeated with pull points from 6 to 15 min every minute.

Enzyme inhibition: As the intermediate product LHF is of interest slowing down or stopping the digestion progress would simplify the procedure. In this experiment the influence of cooling the digestion mixture to rt resp. 5°C, acidifying to pH 5.5 and/or addition of cOmplete™ protease inhibitor was evaluated. As extreme pH changes (e.g. by spiking TFA) might lead to undesired modifications, the maximum pH drop was set to pH 5.5, where most IgG1 are stable.

20.0 µL mIgG1 were diluted in 60.0 µL DB and 5.00 µL of this pre-dilution, 2.50 µL Kgp 10 U/µL and 10.0 µL fresh reducing agent were diluted in 32.5 µL corresponding DB, digested for 15 min at 37°C/300 rpm, 22°C or 5°C and analyzed by CE-SDS NR.

Cysteine influence: To evaluate the influence of cysteine on the enzyme activity several digestions at different concentration levels, origins and age were tested and

23 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

reaction constants compared. As a rate of reduction also the LC peak, at a relative migration time (RMT) of 0.54, was observed through the digestion.

20.0 µL mIgG1 were diluted in 60.0 µL DB to 50.5 mg/mL and the digestion solution were prepared as described in Table 8 with the corresponding RS, incubated for 15 min at 37°C/300 rmp and analyzed by CE-SDS NR.

Table 8: Dilution table of the cysteine influencing digestion.

Genovis Self-made/fresh Self-made/7d-

µL aged 2.0 mM 0.4 mM 2.0 mM 1.0 mM 0.2 mM 1.0 mM 0.2 mM DB pH 8.0 37.5 41.5 37.5 40.0 42.0 40.0 42.0 mIgG1 5.0 5.0 5.0 5.0 5.0 5.0 5.0 50.5 mg/mL Kgp 2.5 2.5 2.5 2.5 2.5 2.5 2.5 10 U/mL RS 20 mM 5.0 1.0 5.0 2.5 0.5 2.5 0.5 Genovis 2.0 and 0.4 mM and fresh/self-made 2.0 mM were then repeated in a three times larger volume and analyzed after 5, 15, 30, 60 and 120 min.

To evaluate the reaction kinetics of the Kgp digestion in absence of cysteine 9.9 µL mIgG1 (50.5 mg/mL) were diluted in 85.1 µL DB and 5.0 µL Kgp (10 U/µL) were added. Afterwards, the digestion mixture was incubated at 37°C/300 rpm for 15, 30 and 60 min and analyzed by CE-SDS NR.

pH influence: To be able to evaluate the impact of the pH and better control the digestion rate mIgG1 was digested at pH 7.3 (slightly basic) and 6.0 (slightly acidic). 20.0 µL mIgG1 were diluted in 60.0 µL DB to 50.5 mg/mL and 9.9 μL of this pre- dilution solution were further diluted in 75.1 μL DB pH 7.3 resp. 6.0, 5.0 μL Kgp (10 U/μL) and 10.0 μL RS were added. Afterwards, the digestion mixture was incubated at 37°C/300 rpm for 15, 30 and 60 min and analyzed by CE-SDS NR.

Digestion parameter optimization: To optimally use the reagents included in the kit and digest to 20 – 30% LHF within a reasonable time frame the enzyme content was set ten times lower compared to the standard protocol and the cysteine concentration was reduced to 1.25 mM. pH 6.0 and 8.0 were tested on those conditions. 20.0 µL mIgG1 solution were diluted in 60.0 µL DB to 50 mg/mL and 19.8 μL of this pre-dilution solution were diluted in 157.7 μL DB at pH 6.0 resp. 8.0, 10 μL Kgp solution (10 U/μL) and 12.5 μL RS were added, incubated at 37°C/300 rpm for 5, 15, 30, 60 and 120 min and analyzed by CE-SDS NR.

24 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

Fc isolation: To confirm the identity of the Fc peak (RMT = 0.73) a Protein L affinity chromatography (AC) was performed. 20.0 µL mIgG1 were diluted in 60.0 µL DB to

50.5 mg/mL, 24.8 μL of this pre-dilution solution were further diluted in 187.7 μL DB, 12.5 μL Kgp (10 U/μL) and 25 μL RS were added and incubated for 2 h at 37°C/300 rpm. After full digestion a sample of 24 µL was taken for CE-SDS NR, the remaining volume was loaded on a Protein L AC pre-equilibrated in binding buffer (BB) and eluted with elution buffer (EB) according to Table 9. Acidic elution fractions were neutralized with 60 µL 1 M tris pH 8.0 per milliliter fraction.

Table 9: Elution of Fc isolation.

Number of Volume in mL Buffer Fraction Fraction

steps volume in mL name 1 1 BB 1.2 Flow through 4 1 BB 4 x 1 Wash 1 – 4 5 1 EB pH 3.5 5 x 1 Elution 1 – 5 The content of each fraction was determined by Nanodrop. The rinsing fraction was concentrated using a 0.5 mL 10K-centrifugal filter, the concentration was determined and the purity analyzed by CE-SDS NR.

SEC screening

To generate material for the preliminary trials 4.00 mg mIgG1 were digested according to the optimized conditions and quenched after 30 min with 84 μL

trifluoroacetic acid 0.5%. Number of columns: To evaluate the optimal number of serially connected columns required to reach the needed resolution, 40 μg mIgG1 were injected on two and three columns in series.

Column load: To evaluate the maximal injection amount without losing resolution 40, 120, 200, 300 and 400 µg mIgG1 were loaded on three columns in series.

Fractionation range: To evaluate an appropriate range to collect desired fractions 20 runs of 40 µg mIgG1 each were injected on three columns in series and collected according to Table 10. Collected fractions were pooled, concentrated with 15 and 0.5 mL 10K-centrifugal filters and analyzed by Nanodrop and CE-SDS NR.

Table 10: SEC fractionation range.

Fragment name Fraction range in min. IgG 59.0 – 61.0 LHF 63.5 – 66.0 Fab (+ Fc) 71.5 – 75.0

25 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

Protein L AC and SEC loading buffer evaluation: To generate fragments free of Fc and to stop further fragmentation the digestion was repeated with 8.00 mg mIgG1.

After 30 min digestion, the mixture was loaded on a Protein L AC column and eluted as described in Table 11. Acidic elution fractions were neutralized with 60 µL 1 M tris pH 8.0 per milliliter fraction.

Table 11: Elution of Protein L AC.

Number of Volume in mL Buffer Fraction Fraction steps volume in mL name 1 1 BB 1.8 Flow through 4 1 BB 4 Wash 8 1 EB pH 2.0 8 Elution The concentration of each fraction was measured. Elution fraction 2 – 8 were pooled, concentrated with a 15 mL 10K-centrifugal filter to approximately 1 mL and analyzed by Nanodrop. 60 µL (413 µg) were analyzed by SEC.

The sample was then buffer exchanged to SEC mobile phase with a 0.5 mL 10K- centrifugal filter, rinsed and diluted with 2 x 500 µL mobile phase. Afterwards, the content was determined. 8 fraction collection runs of 85 µL (8 x 406 µg) were performed. The collected fractions were pooled, concentrated and buffer exchanged into formulation buffer in a 4 mL 10K-centrifugal filter to approximately 150 µL sample. The fractions were analyzed by Nanodrop and CE-SDS NR.

Sample stability: To evaluate the sample stability in mobile phase the injection sample was analyzed before and after fraction collection (T1d) by CE-SDS NR.

Digestion reproducibility and fragment isolation: To evaluate the reproducibility the digestion was repeated with 4.00 mg mIgG1. After 30 min a sample was analyzed by CE-SDS NR. The remaining solution was loaded on a Protein L AC column and eluted as described in Table 12. Acidic elution fractions were neutralized with 60 µL 1 M tris pH 8.0 per milliliter fraction.

Table 12: Elution of Protein L AC for reproducibility.

Number of Volume in mL Buffer Fraction Fraction steps volume in mL name 1 1 BB 1.4 Flow through 4 1 BB 4 Wash 10 1 EB pH 2.0 10 Elution The eluate was concentrated and buffer exchanged to approximately 150 µL and rinsed with 250 µL formulation buffer. The concentration was measured and the samples diluted with additional 800 µL formulation buffer to 3.8 mg/mL. 7 fraction collection runs of 100 µL (7 x 380 µg) were performed. The fractions were pooled,

26 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

concentrated and buffer exchanged into formulation buffer in a 4 and 0.5 mL 10K- centrifugal filter to approximately 42 µL and rinsed with 2 x 10 µL formulation buffer. The fractions were analyzed by Nanodrop and CE-SDS NR.

Digestion progress of IgdE

During the time of this master’s thesis Genovis launched their IgdE product, called FabALACTICA™. Its cleavage pattern is the same as Kgps, apart from the cleavage site shift of one amino acid. Compared the GingisKHAN®, FabALACTICA™ is superior in regard to digestion at neutral pH and absence of reducing agent. However it has a lower digestion rate. In this experiment it was evaluated if IgdE is suitable for LHF generation.

10.0 µL mIgG1 solution was pre-diluted in 40.5 µL IgdE DB to 40.0 mg/mL, 5.0 µL of this pre-dilution solution (200 µg) were further diluted in 10.0 µL DB to 10 mg/mL and 5.0 µL IgdE 40 U/µL (200 U) were added, incubated at 37°C/300 rpm for 3 h and analyzed by CE-SDS NR.

This procedure was repeated with 800 µg mIgG1 and analyzed after 30, 50, 60 and 90 min and a third time for time points at 5, 10, 15, 20 and 25 min.

3.5.2 F(ab’)2

Dig. Digestion IPC In-process control AC Affinity chromatography

BE Buffer exchange

Control Purity analysis

Figure 31: Flow chart of the F(ab)2 generation and isolation.

IdeS digestion and Protein L AC pH 3.5 elution

The IdeS digestion was performed according to the Genovis protocol. 4.95 µL mIgG1

were diluted with 80.0 µL DB and 15.0 µL IdeS (66.7 U/mL) were added. The solution was incubated at 37°C/300 rpm for 30 min. After the digestion a sample of 18.0 µL was taken and analyzed by CE-SDS NR.

Afterwards the Protein L AC was performed according to GE Healthcare Instructions 29-0078-77 AC [69]. The digestion mixture was loaded on the column and eluted as

27 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

listed in Table 13. Acidic elution fractions were neutralized with 60 µL 1 M tris pH 8 per milliliter buffer.

Table 13: Elution of Protein L AC after IdeS digestion.

Number of Volume in mL Buffer Fraction Fraction

steps volume in mL name 1 0.6 BB 0.68 Flow through 5 1.0 BB 5 x 1 Wash 5 1.0 EB 5 x 1 Elution The content of each fraction was determined by Nanodrop. Flow through and wash fraction 1 were pooled, concentrated with 4 and 0.5 mL 10K-centrifugal filter to approximately 20 µL, rinsed with 20 µL formulation buffer and analyzed by CE-SDS NR. All elution fraction were pooled, concentrated with 4 and 0.5 mL 30K-centrifugal filter and rinsed with 2 x 10 µL formulation buffer. Concentration of both samples were determined.

Over-digestion screening and Protein L AC pH 2.0 – 3.0 elution

To check whether the digestion is complete after 30 min and exclude over-digestion,

the digestion time was extended to 60 min. To evaluate the correct pH value to elute

the F(ab’)2 fragment from the column a stepwise elution gradient of pH 3.0, 2.5 and 2.0 was applied.

The digestion conditions are identical to the digestion before. After 60 min digestion a sample of 18.0 µL was taken for CE-SDS NR and the remaining volume was loaded on the Protein L column, washed and eluted according to Table 14.

Table 14: Protein L AC purification table.

Number of Volume in mL Buffer Fraction Fraction name steps volume in mL 2 (vial rinse) 0.45 BB 1 Flow through 4 1.0 BB 4 Wash 8 0.25 EB pH 3.0 5 pH 3.0 elution 3 1.0 8 0.25 EB pH 2.5 5 pH 2.5 elution 3 1.0 8 0.25 EB pH 2.0 5 pH 2.0 elution 3 1.0 The protein content of each fraction was determined. The flow through, wash and all fractions of the identical EB (pH 2.5 and 2.0) were pooled individually, concentrated with a 4 and 0.5 mL 10K-centrifugal filter and each sample was analyzed by Nanodrop and CE-SDS NR.

28 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

FragIT™ kit digestion and fragment isolation

For simple enzyme removal and F(ab’)2 isolation Genovis offers the FragIT™ kit,

containing agarose-immobilized IdeS and a CaptureSelect™ column with an immobilized 13 kDa lama antibody fragment [70]. The digestion and fragment isolation were performed according to the FragIT™ kit instructions [70] including Fc/2 elution and all steps for maximal recovery. To evaluate if the columns could be reused the whole procedure was repeated at the same day and after 17 days.

49.5 µL mIgG1 solution (10 mg IgG) were diluted with 450 µL DB to 20 mg/mL, loaded on the digestion column and rotated end-over-end for 30 min at rt. Afterwards the solution was collected through centrifugation at 100 g for 1 min and the column was washed with 2 x 1 mL BB. The content was determined and a sample for CE- SDS NR was taken. The solution was loaded on the CaptureSelect™ column, rotated end-over-end for 30 min at rt, collected by centrifugation at 200 g for 1 min and washed with 2 x 1 mL BB. The Fc/2 was collected through re-suspending with 2 x 1 mL CaptureSelect™ EB and collecting at 200 g resp. 1000 g at the final step.

Afterwards, the fractions (4.5 mL F(ab’)2 and 2.5 mL Fc/2-fraction) were concentrated and buffer exchanged to formulation buffer with 4 and 0.5 mL 10K- centrifugal filter to approximately 42 µL and rinsed with 3 x 20 µL. Samples were analyzed by Nanodrop and CE-SDS NR. This procedure was repeated after 150 min and 17 days.

3.5.3 Chitobiose cleavage

Dig. Digestion IPC In-process control AC Affinity chromatography

BE Buffer exchange

Control Purity analysis

Figure 32: Flow chart of the EndoS2 digestion and fragment isolation.

EndoS2 digestion

The EndoS2 digestion was performed according to the Genovis protocol [24] at

20 mg/mL. For stability reasons the pH of the DB was set to 6.5.

29 ZHAW LSFM, 2017 Methods and materials Martin Alexander Rappo

5.90 µL mIgG1 (1.19 mg) were diluted in 24.1 µL DB and 30.0 µL EndoS2 40 U/µL (1’200 U) added. The solution was incubated at 37°C/300 rpm for 30 min. After the digestion a sample of 9.0 µL analyzed by CE-SDS red. The remaining digestion mixture was buffer exchanged with a 100 K-spin filter, the membrane was rinsed with 2 x 20 µL fresh formulation buffer and the sample was analyzed by Nanodrop and CE-SDS red.

Immobilized GlycINATOR®

To ensure complete enzyme removal immobilized EndoS2 (immobilized

® GlycINATOR MidiSpin) is used for digestion. The digestion was performed according the Genovis standard protocol for A0-GL6-100 Version 17.1.1 [71] including steps for maximum recovery.

49.5 µL mIgG1 (10 mg) were diluted in 450.5 µL DB, loaded on the equilibrated column and rotated end-over-end for 30 min at rt. Afterwards, the solution was collected through centrifugation at 100 g for 1 min and the column was washed with 2 x 1 mL BB. All fractions were pooled, concentrated and buffer exchanged into formulation buffer with 4 and 0.5 mL 10K-centrifugal filters to approximately 42 µL. The filter was rinsed with 2 x 20 µL formulation buffer and the sample was analyzed by Nanodrop and CE-SDS red.

3.6 Ecology, safety and disposal

Standard laboratory safety equipment consisting of lab coat and protective goggles were used during all tasks. During acid or base handling regular nitrile and during handling of organic solvents butyl gloves were used. Contaminated solutions and consumables were discarded in the chemical waste container. Powdered SDS and 2-mercaptoethanol were only handled under the hood, as SDS creates irritating dust and 2-mercaptoethanol creates toxic fumes and may be fatal in contact with skin [72].

30 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

4 Results and discussion

4.1.1 Fab, Fc and LHF

Stability evaluation in digestion buffer with and without reducing solution

Deamidation stability: A negligible deamidation rate of < 0.5% was detected within the two hours in DB for both temperatures. However after 24 h an increase of 0.7

and 2.6% was detected for rt and 37°C, respectively. The chromatogram is shown in the appendix on p. VII. Neither fragmentation nor aggregation was detected by CE- SDS NR during the measured time frame.

Stability in digestion buffer

Purity of acidic variants by CEX Purity of main by CE-SDS NR 12.5 100 12 99.5 11.5 99 11 98.5 10.5 98 10 97.5 9.5 97

9 96.5 Purity(Main) Purity(Main) in A% Purity(Acidic) Purity(Acidic) in A% 8.5 96 8 95.5 7.5 95 0 4 8 12 16 20 24 0 4 8 12 16 20 24 Time in h Time in h

Figure 33: Purity trend of the stability in digestion buffer Figure 34: Purity trend of the stability in digestion buffer by CEX (deamidation). by and CE-SDS NR (fragmentation/aggregation).

Reduction stability: The relative area increase of the sum of reduced species (L, H, HL, HH, and HHL) after 2 h at 37°C is below 0.1% and therefore consider non- critical. The electropherogram overlay is displayed in Figure 35.

31 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

CE-SDS NR electropherogram - Digestion buffer stability with reducing agent

T0 T120 min Main

HHL

Absorbance in mAb L H HL HH

7 9 11 13 15 17 19 Time in min.

Figure 35: CE-SDS NR electropherogram of the mIgG1 stability in digestion buffer with reducing agent.

Digestion progress of Kgp and influence factors

Enzyme ratio: The consistency of the CE-SDS NR double injection of the digestion sample after 5 min proves, that the digestion stops in presence of 1% SDS. The digestion according to the Genovis protocol is almost complete after 5 min (appendix p. VII) and is therefore not suitable for the generation of LHF. With the adjusted enzyme-protein ratio of 1:10 a maximum LHF concentration of 24 area% is reached after 14 min and the digestion is complete after 2 h. Linear regression of the kinetic

ln([A]/[A]0) VS. time-plot, displayed in the appendix on p. VIII shows a coefficient of determination (R2) of 0.9683. This high value is an indication of first order-like behavior [66]. A comparison of the measured and the simulated values, presented in Figure 36, shows a good overlap of the IgG1 and LHF curve. Fc and Fab concentrations are slightly overestimated by the simulation. As the generation of LHF has priority over Fc and Fab this model is considered fit for purpose.

32 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

1:10 enzyme:protein digestion progress - measured VS. calculated 100% 90% 80% IgG1(measured) 70% IgG1(calculated) 60% LHF (measured) 50% LHF (calculated) 40%

Purity Purity in A% Fc (measured) 30% Fc (calculated) 20% Fab (measured) 10% Fab (calculated) 0% 0 15 30 45 60 75 90 105 120 Time in min

Figure 36: CE-SDS NR purity progress of the 1:10 Kgp-to-IgG digestion and comparison of the simulated values.

Enzyme inhibition: An evaluation of the reaction constant k1 (see Figure 37) and

tmax(LHF) (see Figure 38) revealed following inhibition potential:

Cooling > Acidify to pH 5.5 > cOmplete™

Inhibitory factor influence

Kinetic constant Time of maximal LHF 0.12 conentration R² = 0.9946 120 0.10 100 0.08 R² = 0.9999 80 0.06 R² = 0.9654 60

0.04 in min(LHF)

max

t Reaction Reaction constant k1 0.02 20

0.00 0 5 15 25 35 5 15 25 35 Temperature / °C Temperature / °C

Figure 37: Reaction constant k1 in dependency of Figure 38: Calculated time point of maximal LHF temperature for tested inhibitory factors (standard concentration in dependency of temperature for tested digestion, cOmplete™ and pH 5.5). inhibitory factors (standard digestion, cOmplete™ and pH 5.5).

33 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

The cOmplete™ protease inhibitor cocktail will be omitted for further studies, as the inhibition potential is low. Additionally, as the composition is not clearly defined tracking during fragment isolation and full removal cannot be ensured. For further studies cooling to 5°C potentially in combination with acidifying and enzyme removal for minimal molecule impact will be applied.

Cysteine influence: The comparison between digestions a different cysteine concentrations, shown in Figure 39 and Figure 40, revealed a high influence of cysteine on the Kgp digestion rate. The Kgp digestion rate is drastically reduced resp. the digestion time is increased in absence of cysteine.

Cysteine concentration influence

Kinetic constant Time point of the maximal LHF 0.20 concentration 0.18 250 0.16 0.14 200 0.12 in min

0.10 max 150 0.08 100 0.06

Kinetic Kinetic constant k1 0.04 50

0.02 Timeof[LHF] 0.00 0

Cysteine concentration in mM Cysteine concentration in mM

Figure 39: Kinetic constants of the Kgp digestion for Figure 40: Calculated time point of maximal LHF different cysteine concentration of Genovis, self- concentration of the Kgp digestion for different cysteine made/fresh and self-made/aged RS. concentration of Genovis, self-made/fresh and self- made/aged RS.

The self-made cysteine solution had the highest digestion rate resp. lowest digestion time for all concentrations compared to other conditions. The digestion using aged self-made cysteine solution was slower than freshly made RS for all concentrations. The highest digestion rate and lowest digestion time was determined at a concentration of 1 mM, even though similar results were obtained for a concentration of 2 mM cysteine. A possible explanation for the lower kinetic constant of Genovis’ RS compared to the self-made might be small differences regarding final concentration. The cysteine in aged RS is most probably partly oxidized and not able to provide the optimal reducing environment anymore.

34 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

The purity progress of the peak at RMT 0.54 through the digestion time, displayed in Figure 41, shows a comparable pattern for all tested cysteine concentrations. The purity initially increases and later decreases again. As expected, the rate of reduction increases with increasing cysteine content, but is throughout low. All experiments using the RS of Genovis show a purity maximum at 30 min, whereas the self-made RS shows its maximum after 60 min, which is significant higher compared to the Genovis RS at equivalent concentration. This is an indication that this peak does not only contain LC but also another unidentified species, which is formed and degraded during the digestion.

Purity of the peak at a RMT = 0.54 through the digestion course at different cysteine conentrations 9% 8% 7% 6% 5% Genovis 4 mM 4% Genovis 2 mM

Purity Purity in A% 3% Genovis 1 mM 2% Self-made 2 mM 1% 0% 0 15 30 45 60 75 90 105 120 Time in min

Figure 41: Progress of the 25 kDa-peak formation during the IgG1 Kgp digestion at different cysteine concentrations.

Therefore the cysteine concentration is considered crucial for the digestion progress and Genovis and self-made RS do not deliver equal results.

pH influence: The experiment showed that the pH has a large impact on the digestion kinetic. At pH 8.0 the highest and at pH 6.0 lowest kinetic constant was measured. This is visualized in Figure 42 and Figure 43.

35 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

pH influence Kinetic constant Time point of the maximal LHF 0.18 concentration 0.16 35 0.14 30 0.12

in in min 25 0.1 max 20 0.08 15 0.06 10 Kinetic Kinetic constant k1 0.04

0.02 Timeof[LHF] 5 0 0 6 6.5 7 7.5 8 6 6.5 7 7.5 8 pH pH

Figure 42: Kinetic constants of the Kgp digestion at Figure 43: Calculated time point of maximal LHF different pH. concentration of the Kgp digestion at different pH.

Digestion parameter adjustment: It was observed that upon reconstitution of the cysteine according to the Genovis protocol, a suspension was obtained. The cysteine only fully dissolved after further dilution for digestion. The purity of IgG, LHF, Fc and Fab at pH 6.0 and 8.0 are visualized in Figure 44 and Figure 45, respectively. The digestion at pH 6.0 is relatively slow and the maximum concentration of LHF is not reached until the upper time limitation of 120 min. The digestion at pH 8.0 on the other hand shows a maximum LHF concentration after 30 min followed by a shallow decrease, allowing the scientist to perform the digestion in a reasonable time frame.

36 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

Digestion parameter adjustment

pH 6.0 pH 8.0 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50%

40% 40% Purity Purity in A% Purity Purity in A% 30% 30% 20% 20% 10% 10% 0% 0% 0 15 30 45 60 75 90 105 120 0 15 30 45 60 75 90 105 120 Time in min Time in min

Figure 44: Digestion course of the parameter adjusted Figure 45: Digestion course of the parameter adjusted Kgp digestion at pH 6.0. Kgp digestion at pH 8.0.

Fc isolation: The Protein L AC successfully bound the Fab and the Fc could be collected with a recovery of 47%. Later Fab elution, displayed in Figure 46, was minimal with a recovery of 20%. A possible explanation is, that the pH of the EB was not low enough to reverse the binding.

Protein content of the Protein L affinity column fractions 0.30

0.25

0.20

0.15

0.10

0.05 Concentration Concentration in mg/mL

0.00

Step

Figure 46: Protein content of the Protein L affinity chromatography by Nanodrop.

37 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

The Fc could successfully be isolated with a final concentration of 3.43 mg/mL by Protein L AC and the peak at RMT 0.73 in CE-SDS NR could be assigned. The corresponding electropherogram is shown in Figure 47.

CE-SDS NR electropherogram before and after Protein L AC

Kgp digestion Absorbance in mAU Protein L flow through

6 7 8 9 10 11 12 13 14 15 16 Time in min

Figure 47: CE-SDS NR electropherogram of the Kgp digestion and concentrated Protein L flow through.

SEC screening

Number of columns: Two SEC columns in series are capable of separating IgG, LHF and Fab. Three columns in series prolong the run time from 55 to 85 min, increase the resolution and an additional peak between LHF and Fab could be separated. Baseline separated is not reached in both cases. Corresponding chromatograms are shown in Figure 48.

38 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

SEC chromatogram of the Kgp digest - 2 VS. 3 columns

2 columns 3 columns

IgG1

Fab

IgG1 LHF

LHF

Fab Absorbance in mAU

0 10 20 30 40 50 60 70 80 Time in min

Figure 48: SEC chromatogram overlay of 2 and 3 columns in series.

Column volume: Testing column loads from 40 to 400 μg mIgG1 showed no significant loss of peak resolution. The SEC chromatogram overlay is displayed in Figure 49.

SEC chromatogram - Increasing column load

40 µg 120 µg 200 µg 300 µg 400 µg Absorbance in mAU

55 60 65 70 75 80 85 Time in min

Figure 49: SEC chromatogram of the injection amount evaluation.

Fractionation range: The chosen fractionation range was sufficient to isolate the desired fragments. Importantly, the chromatographic overlays (Figure 50) revealed, that the TFA spiking did not fully quench the digestion and the sample did fragmentation continued at 5°C (shown by decreasing peaks at retention time 59.6

39 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

(IgG) and 64.5 min (LHF) as well as an increase in the Fab peak at retention time 72 min).

Figure 50: SEC chromatogram overlay of the fraction collection trial of 20 times 40 μg injection.

A summary of the isolated fragment is listed in Table 15.

Table 15: Summary of isolated fragments during SEC trials.

Fragment Volume in Conc. In Mass in µg Recovery Purity in µL mg/mL in % A% IgG1 120 0.10 11 4 86 LHF 120 0.18 22 14 75 Fab+Fc 120 0.81 64+32 45+30 98 CE-SDS NR electropherogram - Purity of isolated fragments

IgG LHF Fab+Fc 71% 75% 86%

27% Absorbance in mAU

9 10 11 12 13 14 15 16 Time in min.

Figure 51: CE-SDS NR overlay of the isolated IgG, LHF and Fab+Fc fractions.

40 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

Protein L AC and SEC loading buffer evaluation: The Protein L AC could successfully remove the Fc part and the enzyme (see appendix p. VIII) and stop

further digestion. As a trade-off, a small amount of IgG, LHF and Fab was lost during the wash steps. The intact IgG purity by CE-SDS NR was higher than expected. Reason for this could possibly be a slower digestion rate and imply issues regarding reproducibility. Telquel injection on the SEC showed a different chromatographic pattern, displayed in Figure 52, than previous injections without prior Protein L AC. A possible explanation for this behavior is the different buffer composition influencing the hydrodynamic radius.

Comparison: Fraction collection after digest and after Protein L AC

Without Protein L AC With Protein L AC IgG fraction

LHF fractio Fab fraction Absorbance in mAU

43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 Time in min

Figure 52: SEC chromatogram of the Kgp digestion fragment isolation with and without Protein L AC.

Prior buffer exchange into SEC mobile phase improved pattern resemblance on the SEC column, shown in Figure 53. Peaks before the IgG peak indicate sample instability of mixed sample but also isolated fragments. The lower UV trace of one of the sample injections can be explained by a smaller injection volume. A summary of isolated fragments is listed in Table 16. Reason for the low LHF purity is the relatively high IgG content in the initial sample tailing into the LHF fraction.

Table 16: Summary of isolated fragments during SEC injection buffer evaluation.

Fragment Volume in Conc. In Mass in µg Recovery Purity in µL mg/mL in % A% IgG1 130 4.55 591 89 98 LHF 110 1.29 142 21 45 Fab 110 1.24 136 15 95

41 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

Kgp digestion after Protein L AC and desalting: Fraction collection

Injection 1 Injection 2 Injection 3 Injection 4 Injection 5

Injection 6 Injection 7 IgG fraction LHF fraction Fab fraction Absorbance in mAU

35 40 45 50 55 60 65 70 75 80 85 Time in min

Figure 53: SEC chromatogram overlay of the fragment isolation after Kgp digestion, Protein L AC and buffer exchange in mobile phase.

Sample stability: The CE-SDS NR electropherogram of the SEC injection sample before and after fraction collection, displayed in Figure 54, shows no significant changes. A possible explanation are reversible aggregates, which dissociate in CE- SDS NR sample buffer.

CE-SDS NR electropherogram - Injection sample stability in SEC mobile phase at 5°C

Before fraction collection After fraction collection Absorbance in mAU

9 10 11 12 13 14 15 16 17 18 19 Time in min

Figure 54: CE-SDS NR electropherogram of initial and one-day aged injection sample.

Digestion reproducibility and fragment isolation: CE-SDS NR comparison of two digestions with identical parameters but different GingisKHAN® batches (see Figure

42 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

55) showed inconsistent digestion progress. A possible explanation for this are differences of the RS, as a different solubility was overserved during reconstitution.

CE-SDS NR electropherogram - Comparison of two 30 min Kgp digestions

Digest 1 Digest 2 Absorbance in mAU

9 10 11 12 13 14 15 16 17 Time in min

Figure 55: CE-SDS NR electropherogram overlay of two Kgp digestions with identical parameter apart from different GingisKHAN® batches.

Buffer exchange into formulation buffer improved the sample stability and less peaks before IgG were observed.

SEC chromatogram - Kgp digestion after Protein L AC and buffer exchange in formulation buffer: Fraction collection

Injection 1 Injection 2 Injection 3 Injection 4 Injection 5

Injection 6 IgG fraction LHF fraction Fab fraction Absorbance in mAU

35 40 45 50 55 60 65 70 75 80 85 Time in min

Figure 56: SEC chromatogram overlay of the fragment isolation after Kgp digestion, Protein L AC and buffer exchange in formulation buffer.

43 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

A summary of isolated fragments is listed below.

Table 17: Summary of isolated fragments during SEC fraction collection.

Fragment Volume in Conc. In Mass in µg Recovery Purity in µL mg/mL in % A% IgG1 20 6.21 124 4 97 LHF 20 2.24 45 9 41 Fab 20 2.32 46 15 82

Digestion progress of IgdE

The IgdE digestion progress, displayed in Figure 57, showed a much lower digestion time required than the 16 –18 h suggested in the Genovis protocol. The maximum LHF concentration of 29% was detected after 10 min and would even allow reduction of the enzyme-to-protein ratio to slow down the digestion. The higher intermediate concentration, the poly-his tag for simple enzyme removal and mild digestion conditions make IgdE superior to Kgp for the LHF generation.

Digestion progress of the IgdE digestion 100% 90% 80% 70% 60% IgG 50% LHF 40%

Purity Purity in A% Fc 30% Fab 20% 10% 0% 0 15 30 45 60 75 90 105 120 135 150 165 180 Time in min

Figure 57: Digestion progress of IgdE by CE-SDS NR.

4.1.2 F(ab’)2

IdeS digestion and Protein L AC pH 3.5 elution

The IdeS digestion was complete after 30 min. Fc/2, IdeS and some other, minor

impurities could successfully be removed by the Protein L AC. The F(ab’)2 could not be eluted with EB at pH 3.5.

44 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

Over digestion screening and Protein L AC pH 2.0 – 3.0 elution

After 1 h of digestion no sign of over-digestion could be detected by CE-SDS NR. Elution at pH 3.0 was very slow, at pH 2.5 moderate and at pH 2.0 acceptable. The electropherogram is displayed in the appendix on p. IX.

Protein L AC elution at pH 3.0 - 2.0 0.1 0.09 0.08 0.07 0.06 0.05 pH 3.0 0.04 pH 2.5 0.03 pH 2.0

0.02 Concentration Concentration in mg/mL 0.01 0

Elution buffer volume in mL

Figure 58: Elution profile of Protein L AC at pH 3.0, 2.5 and 2.0.

FragIT™ kit digestion and fragment isolation

The on-column IdeS digestion was complete after 30 min and the Fc/2 could successfully be removed with the CaptureSelect™ column, as shown in Figure 59.

CE-SDS NR electropherogram - FabIT kit digestion and purification

Undigested

Digested

F(ab)2 Absorbance in mAU

Fc/2 7 8 9 10 11 12 13 14 15 16 Time in min

Figure 59: CE-SDS NR electropherogram of the IdeS digestion and isolated products.

45 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

Initial and second digestion at the same day show the same digestion rate. The digestion after 17 days showed a loss in activity and is after 30 min not complete. No differences of digestion products was observed. A summary of the isolated fragments is listed in Table 18.

Table 18: Summary of isolated fragments for FragIT™ kit digestion and fragment isolation.

Run Volume in Conc. in Mass in µg Recovery Purity in

µL mg/mL in % A% 1 100 31.65 3’165 48 91 2 100 36.12 3’612 55 96 3 100 12.71 1271 19 97 CE-SDS NR electropherogram - Overlay of FragIT™ digestion at different time points

1st run (T0) 2nd run (T150min.) 3rd run (T17d) Absorbance in mAU

7 8 9 10 11 12 13 14 15 Time in min

Figure 60: CE-SDS NR electropherogram overlay of the three FragIT™ digestion.

4.1.3 Chitobiose cleavage

EndoS2 digestion

The digestion according to the Genovis® protocol and the subsequent buffer exchange were successful. The enzyme could not be removed but reduced to below 0.5%. 40 µL chitobiose-cleaved mIgG1 27.06 mg/mL (1’082 µg = 94% yield) could be isolated.

46 ZHAW LSFM, 2017 Results and discussion Martin Alexander Rappo

CE-SDS red. electropherogram of the soluble EndoS2 digestion and buffer exchange

Reference Digestion T30min Buffer exchange Absorbance in mAU

15 17 19 21 23 25 27 29 31 33 35 Time in min.

Figure 61: CE-SDS red. electropherogram of the EndoS2 digestion and buffer exchange.

Immobilized GlycINATOR®

The Genovis® digestion procedure for the immobilized GlycINATOR® was successfully applied to mIgG1. The digestion product was be concentrated and buffer exchanged into formulation buffer. 62 µL chitobiose-cleaved mIgG1 at 125 mg/mL (7’750 µg = 80% yield) were isolated.

47 ZHAW LSFM, 2017 Conclusion Martin Alexander Rappo

5 Conclusion

The influence of the undesired modifications, deamidation and reduction during digestion where assessed and are considered minimal (> 0.5%) within 2 The Kgp digestion of the mIgG1 using GingisKHAN® was successful and the LHF could be detected as intermediate by CE- SDS NR after reducing the enzyme-to-protein ratio to 1:10. Cysteine concentration and pH of the DB was classified as a crucial influence parameter for the digestion kinetic. Through digestion parameter optimization for LHF isolation the time for the maximal LHF was adjusted to 30 min at 37°C, which is considered an appropriate time frame to handle. Of the evaluated digestion inhibition parameter (temperature, pH and cOmplete™ inhibitor cocktail) nothing fully inhibited but slowed down further fragmentation. Inhibition potential is in decreasing order: cooling to 5°C, acidify to pH 5.5 and adding cOmplete™. Therefore enzyme removal by AC is proposed to stop further digestion. Through Protein L AC the Fc could be isolated and following elution at pH 2.0 the IgG, LHF and Fab mixture could be eluted. Small quantities could be separated by SEC with a purity of > 75%. Scale-up and reproducibility experiments revealed issues regarding inconsistent digestion progresses for different batches. Slow digestion led finally to large IgG contamination of the LHF fraction. This digestion kinetic variations probably origin from RS differences. Experiments on the recently launched FabALACTICA™ (IgdE) showed promising results with an acceptable digestion time, mild digestion conditions, less influencing factors and a poly-his tag on the enzyme for simple removal and digestion stop. Also the outlook on the immobilized IgdE, which is according to Genovis® supposed to be launched in fall 2017 makes IgdE and better alternative for LHF generation than Kgp.

Digestion with FabRICATOR® (soluble IdeS) was complete after 30 min and Fc/2 was removed by Protein L AC. The F(ab‘)2 was successfully generated by applying standard FragIT™ (immobilized IdeS) procedure and the desired fragment was isolated by CaptureSelect™ without exposure to acidic solution. The maximal recovery was 55% and purity 96% and no undesired modifications were detected. The procedure was repeated with identical outcome at the same day using the same mab and digestion kit.

® The chitobiose cleavage by GlyCINATOR (soluble EndoS2) was complete after 30 min at 37°C but a simple 100K-centrifugal filtration was not sufficient to fully remove the enzyme. Immobilized GlyCINATOR® provided a full digestion after 30 min at rt and full enzyme removal with a yield of 80%.

48 ZHAW LSFM, 2017 Outlook Martin Alexander Rappo

6 Outlook

Following experiments should include identification and in-depth analysis of undesired modification by mass spectrometry.

Also other types of IgG1, mutants, other applicable subclasses or proteins should be tested on the implemented methods.

As soon as the immobilized FabALACTICA™ (IgdE) is launched, digestion progress and applicability for the LHF generation should be evaluated. Important is also a reproducibility and batch-to-batch comparison.

Isolated digestion products have to be tested for regarding CQA. This could include: cell bioassay for fragments and higher order structure analysis by differential scanning calorimetry and circular dichroism in the far UV for the chitobiose chelated product.

49 ZHAW LSFM, 2017 References Martin Alexander Rappo

7 References

[1] R.O. Esquivel, M. Molina-Espiŕ itu, F. Salas, C. Soriano, C. Barrientos, J.S. Dehesa, J.A. Dobado, Decoding the building bloks of life from the perspective of quantum information, Licens. InTech. (2013) 641–669. doi:http://dx.doi.org/10.5772/55160 1. [2] J.K.H. Liu, The history of monoclonal antibody development - Progress, remaining challenges and future innovations, Ann. Med. Surg. 3 (2014) 113–116. doi:10.1016/j.amsu.2014.09.001. [3] A.L. Nelson, E. Dhimolea, Development trends for human monoclonal antibody therapeutics, Nat. Rev. Drug Discov. 9 (2010) 767–774. doi:10.1038/nrd3229. [4] H.M. Shepard, G.L. Phillips, C.D. Thanos, M. Feldmann, Developments in therapy with monoclonal antibodies and related proteins, Clin. Med. J. R. Coll. Physicians London. 17 (2017) 220–232. doi:10.7861/clinmedicine.17-3-220. [5] D.M. Ecker, S.D. Jones, H.L. Levine, The therapeutic monoclonal antibody market., MAbs. 7 (2015) 9–14. doi:10.4161/19420862.2015.989042. [6] ICH, ICH Quality Guidelines, (2017). http://www.ich.org/products/guidelines/quality/article/quality-guidelines.html (accessed August 2, 2017). [7] U.S. Department of Health and Human Services Food and Drug Administration, Q8, Q9, & Q10 Questions and Answers -- Appendix: Q&As from Training Sessions (Q8, Q9, & Q10 Points to Consider), (2012). https://www.fda.gov/drugs/guidancecomplianceregulatoryinformation/guidances/ucm3 13087.htm (accessed August 2, 2017). [8] EMA, EMA guideline Q11 - Development and Manufacture of Drug Substances (chemical entities and biotechnological / biological entities), 2011. doi:10.5639/gabij.2012.0103.025. [9] N. Alt, T.Y. Zhang, P. Motchnik, R. Taticek, V. Quarmby, T. Schlothauer, H. Beck, T. Emrich, R.J. Harris, Determination of critical quality attributes for monoclonal antibodies using quality by design principles, Biologicals. 44 (2016) 291–305. doi:10.1016/j.biologicals.2016.06.005. [10] ICH, ICH Q8(R2) Pharmaceutical Development, 2009. [11] ICH, ICH Q9- Quality Risk Management, 2005. doi:10.1007/s11095-007-9511-1. [12] ICH, ICH Q10 - Pharmaceutical quality system, 2008. doi:EMEA/CHMP/ICH/214732/2007. [13] ICH, ICH Q11 - Development and Manufacture of Drug Substances (Chemical Entities and Biotechnological/Biological Entities, 2012. doi:10.5639/gabij.2012.0103.025. [14] ICH, ICH Final Concept Paper Q12 - Technical and Regulatory Considerations for Pharmaceutical Product Lifecycle Management, 2014. [15] Novartis Brandlab, Assets, (2017). [16] S. Kim, Quality by Design (QbD) Framework, (2014). https://www.youtube.com/watch?v=XlPlDV8ogyo (accessed September 25, 2017). [17] Antibody Solutions, Potency assays, (2017). https://antibody.com/solutions/by- applications/potency-assays/. [18] J. Carlsson, E. Forssell Aronsson, S.-O. Hietala, T. Stigbrand, J. Tennvall, E.T. Boder, W. Jiang, A. Labeling, I. Absorption, D. Biotransformation, E. General, C. Multiple, C.L. Maynard, C.O. Elson, R.D. Hatton, C.T. Weaver, American Society of Health-System Pharmacists, Introduction to Pharmacokinetics and Pharmacodynamics,

50 ZHAW LSFM, 2017 References Martin Alexander Rappo

Pharmocokinetics and Pharmcodynamics. 1 (2010) 1–18. doi:10.1007/SpringerReference_184421. [19] S. Ait-Oudhia, M.A. Ovacik, D.E. Mager, Systems pharmacology and enhanced pharmacodynamic models for understanding antibody-based drug action and toxicity, MAbs. 20 (2016) 00–00. doi:10.1080/19420862.2016.1238995. [20] WHO, Annex 7 Application of Hazard Analysis and Critical Control Point ( HACCP ) methodology to pharmaceuticals, WHO Tech. Rep. Ser. No. 908, 2003 Annex. (2003) 99–112. [21] Genovis, FabRICATOR instructions, (n.d.). https://www.genovis.com/wp- content/uploads/instructions-fabricator-2000-5000.pdf. [22] Genovis, GingisKHAN instructions, (n.d.). https://www.genovis.com/wp- content/uploads/INSTRUCTIONS_GingisKHAN_17.1.1.pdf. [23] Genovis, FabALACTICA instructions, (n.d.). https://www.genovis.com/wp- content/uploads/INSTRUCTIONS-for-FabALACTICA-2000u_version-17.1.1.pdf. [24] Genovis, GlycINATOR instructions, (n.d.). https://www.genovis.com/wp- content/uploads/INSTRUCTIONS-for-GlycINATOR_2000u_version-17.1.1.pdf. [25] C. Schütt, B. Bröker, Grundwissen Immunologie, 2., Springer, Heidelberg, 2009. [26] S.N. WEI WANG, SATISH SINGH, DAVID L. ZENG, KEVIN KING, Antibody Structure, Instability, and Formulation, J. Pharm. Sci. 96 (2007) 1–26. doi:DOI 10.1002/jps.20727. [27] R. Jefferis, M.-P. Lefranc, Human immunoglobulin allotypes, MAbs. 1 (2009) 1–7. doi:10.4161/mabs.1.4.9122. [28] H. Liu, K. May, Disulfide bond structures of IgG molecules, MAbs. 4 (2012) 17–23. doi:10.4161/mabs.4.1.18347. [29] A.R. Costa, M.E. Rodrigues, M. Henriques, R. Oliveira, J. Azeredo, Glycosylation: impact, control and improvement during therapeutic protein production., Crit. Rev. Biotechnol. 8551 (2013) 1–19. doi:10.3109/07388551.2013.793649. [30] S. Ha, Y. Ou, J. Vlasak, Y. Li, S. Wang, K. Vo, Y. Du, A. MacH, Y. Fang, N. Zhang, Isolation and characterization of IgG1 with asymmetrical Fc glycosylation, Glycobiology. 21 (2011) 1087–1096. doi:10.1093/glycob/cwr047. [31] Boundless, Antibody Proteins and Antigen Binding, (2014). https://www.boundless.com/microbiology/textbooks/boundless-microbiology- textbook/immunology-11/antibodies-141/antibody-proteins-and-antigen-binding-714- 5819/ (accessed August 9, 2017). [32] R.C. Hillig, S. Urlinger, J. Fanghänel, B. Brocks, C. Haenel, Y. Stark, D. Sülzle, D.I. Svergun, S. Baesler, G. Malawski, D. Moosmayer, A. Menrad, M. Schirner, K. Licha, Fab MOR03268 Triggers Absorption Shift of a Diagnostic Dye via Packaging in a Solvent-shielded Fab Dimer Interface, J. Mol. Biol. 377 (2008) 206–219. doi:10.1016/j.jmb.2007.12.071. [33] E.O. Saphire, Crystal Structure of a Neutralizing Human IgG Against HIV-1: A Template for Vaccine Design, Science (80-. ). 293 (2001) 1155–1159. doi:10.1126/science.1061692. [34] G. Vidarsson, G. Dekkers, T. Rispens, IgG subclasses and allotypes: From structure to effector functions, Front. Immunol. 5 (2014) 1–17. doi:10.3389/fimmu.2014.00520. [35] abcam, Antibody structure and isotypes, (2017) 1–4. http://docs.abcam.com/pdf/antibody-guide/antibody-structure-and-isotypes.pdf. [36] C. Spoerry, P. Hessle, M.J. Lewis, L. Paton, J.M. Woof, U. Von Pawel-rammingen,

51 ZHAW LSFM, 2017 References Martin Alexander Rappo

Novel IgG-Degrading Enzymes of the IgdE Protease Family Link Substrate Specificity to Host Tropism of Streptococcus Species, PLoS One. 11 (2016) 1–20. doi:10.1371/journal.pone.0164809. [37] D.L. Chabner, Bruce A.; Longo, Cancer Chemotherapy and Biotherapy: Principles and Practice, 5th ed., Wolters Kluwer: Lipincott Williams & Wilkins, 2011. https://books.google.ch/books?id=0U4aj4GZWCIC&pg=PA468&lpg=PA468&dq=antib odyfragment+%22Fabc%22&source=bl&ots=F4243KCPtq&sig=T- PqR_9hSSqNjr8R4XlhxZe41QA&hl=en&sa=X&ved=0ahUKEwj63fDelsrVAhXMuhQK HV1KD_UQ6AEIOzAC#v=onepage&q&f=false. [38] Thermo Fisher Scientific, Antibody Fragmentation, (n.d.). https://www.thermofisher.com/ch/en/home/life-science/antibodies/antibodies-learning- center/antibodies-resource-library/antibody-methods/antibody-fragmentation.html (accessed August 9, 2017). [39] Absolute Antibody, Antibody Fragments, (n.d.). http://absoluteantibody.com/antibody- resources/antibody-engineering/antibody-fragments/ (accessed August 9, 2017). [40] E. O’Connor, M. Aspelund, F. Bartnik, M. Berge, K. Coughlin, M. Kambarami, D. Spencer, H. Yan, W. Wang, Monoclonal antibody fragment removal mediated by mixed mode resins, J. Chromatogr. A. 1499 (2017) 65–77. doi:10.1016/j.chroma.2017.03.063. [41] IMGT, Organization in domains of an IgG1 immunoglobulin and of its fragments, (2001). http://www.imgt.org/IMGTeducation/Tutorials/IGandBcells/_UK/3Dstructure/Figure1.ht ml (accessed May 2, 2017). [42] R.S. Shochat, Dan; Hansen, Hans J.; Wu, Method for radiolabeling antibody fragments, US5514363 (A) ― 1996-05-07, 1993. https://worldwide.espacenet.com/publicationDetails/biblio?II=0&ND=3&adjacent=true& locale=en_EP&FT=D&date=19960507&CC=US&NR=5514363A&KC=A#. [43] J. Rohrer, Master Course: Comparative Physiology Glycosylation in Eukaryotes, (2017). [44] F. Higel, A. Seidl, F. Sörgel, W. Friess, N-glycosylation heterogeneity and the influence on structure, function and pharmacokinetics of monoclonal antibodies and Fc fusion proteins, Eur. J. Pharm. Biopharm. 100 (2016) 94–100. doi:10.1016/j.ejpb.2016.01.005. [45] P.M. Rudd, R. a Dwek, Glycosylation: heterogeneity and the 3D structure of proteins., Crit. Rev. Biochem. Mol. Biol. 32 (1997) 1–100. doi:10.3109/10409239709085144. [46] K. Zheng, C. Bantog, R. Bayer, The impact of glycosylation on monoclonal antibody conformation and stability, MAbs. 3 (2011) 568–576. doi:10.4161/mabs.3.6.17922. [47] K. Zheng, M. Yarmarkovich, C. Bantog, R. Bayer, T.W. Patapoff, Influence of glycosylation pattern on the molecular properties of monoclonal antibodies, MAbs. 6 (2014) 649–658. doi:10.4161/mabs.28588. [48] MEROPS, Summary for peptidase C66.001: IdeS peptidase, (2017). https://www.ebi.ac.uk/merops/cgi-bin/pepsum?mid=C66.001 (accessed August 11, 2017). [49] C. Spoerry, J. Seele, P. Valentin-Weigand, C.G. Baums, U. von Pawel-Rammingen, Identification and Characterization of IgdE, a Novel IgG-degrading Protease of Streptococcus suis with Unique Specificity for Porcine IgG, Am. Soc. Biochem. Mol. Biol. (2016). doi:10.1074/jbc.M115.711440. [50] K. Wenig, L. Chatwell, U. von Pawel-Rammingen, L. Björck, R. Huber, P. Sondermann, Structure of the streptococcal endopeptidase IdeS, a cysteine

52 ZHAW LSFM, 2017 References Martin Alexander Rappo

proteinase with strict specificity for IgG., Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 17371–6. doi:10.1073/pnas.0407965101. [51] N. Fittipaldi, S.B. Beres, R.J. Olsen, V. Kapur, P.R. Shea, M.E. Watkins, C.C. Cantu, D.R. Laucirica, L. Jenkins, A.R. Flores, M. Lovgren, C. Ardanuy, J. Liares, D.E. Low, G.J. Tyrrell, J.M. Musser, Full-genome dissection of an epidemic of severe invasive disease caused by a hypervirulent, recently emerged clone of group A Streptococcus, Am. J. Pathol. 180 (2012) 1522–1534. doi:10.1016/j.ajpath.2011.12.037. [52] P.P.P. Cleary, Streptococcus pyogenes disease and molecular pathogenesis, HSTalks, USA, 2009. https://hstalks.com/t/1480/streptococcus-pyogenes-disease-and- molecular-patho/. [53] M.W. Robinson, J.P. Dalton, Cysteine Proteases of Pathogenic Organisms, Springer Science+Business Media, New York USA, 2011. doi:10.1007/978-1-4419-8414-2. [54] B. Vincents, A. Guentsch, D. Kostolowska, U. von Pawel-Rammingen, S. Eick, J. Potempa, M. Abrahamson, Cleavage of IgG1 and IgG3 by gingipain K from Porphyromonas gingivalis may compromise host defense in progressive periodontitis, FASEB J. 25 (2011) 3741–3750. doi:10.1096/fj.11-187799. [55] C. Spoerry, Streptococcal Immunoglobulin degrading Enzymes of the IdeS and IgdE Family, 2017. [56] N. Pavloff, P. a Pemberton, J. Potempa, W.A. Chen, R.N. Pike, V. Prochazka, M.C. Kiefer, J. Travis, P.J. Barr, P.J.J.B. Chem, Molecular Cloning and Characterization of Porphyromonas gingivalis Lysine-specific Gingipain, J. Bilogical Chem. 272 (1997) 1595–1600. [57] I. De Diego, F. Veillard, M.N. Sztukowska, T. Guevara, B. Potempa, A. Pomowski, J.A. Huntington, J. Potempa, F.X. Gomis-Rüth, Structure and mechanism of cysteine peptidase gingipain K (Kgp), a major virulence factor of porphyromonas gingivalis, J. Biol. Chem. 289 (2014) 32291–32302. doi:10.1074/jbc.M114.602052. [58] I. de Diego, F. Veillard, M.N. Sztukowska, T. Guevara, B. Potempa, A. Pomowski, J.A. Huntington, J. Potempa, F.X. Gomis-Rüth, Structure and Mechanism of Cysteine Peptidase Gingipain K (Kgp), a Major Virulence Factor of Porphyromonas gingivalis in Periodontitis, J. Biol. Chem. 289 (2014) 32291–32302. doi:10.1074/jbc.M114.602052. [59] D.E. Koshland, K. Hamadani, Proteomics and models for enzyme cooperativity, J. Biol. Chem. 277 (2002) 46841–46844. doi:10.1074/jbc.R200014200. [60] MEROPS, Summary for C113.001, (2017). https://www.ebi.ac.uk/merops/cgi- bin/pepsum?id=C113.001 (accessed August 16, 2017). [61] MEROPS, MER1020437 - IgdE peptidase ({Streptococcus suis}), (2017). https://www.ebi.ac.uk/merops/cgi-bin/aaseq?mernum=MER1020437 (accessed September 25, 2017). [62] J. Sjögren, W.B. Struwe, E.F.J. Cosgrave, P.M. Rudd, M. Stervander, M. Allhorn, A. Hollands, V. Nizet, M. Collin, EndoS 2 is a unique and conserved enzyme of serotype M49 group A Streptococcus that hydrolyses N-linked glycans on IgG and α 1 -acid glycoprotein, Biochem. J. 455 (2013) 107–118. doi:10.1042/BJ20130126. [63] J. Sjögren, W.B. Struwe, E.F.J. Cosgrave, P.M. Rudd, M. Stervander, M. Allhorn, A. Hollands, V. Nizet, M. Collin, endo-beta-N-acetylglucosaminidase [Streptococcus pyogenes], EndoS 2 Is a Unique Conserv. Enzym. Serotype M49 Gr. A Streptococcus That Hydrolyses N-Linked Glycans IgG α 1 -Acid Glycoprotein. (2013). https://www.ncbi.nlm.nih.gov/protein/AGU16859.1 (accessed September 5, 2017). [64] N. Fittipaldi, S.B. Beres, R.J. Olsen, V. Kapur, P.R. Shea, M.E. Watkins, C.C. Cantu, D.R. Laucirica, L. Jenkins, A.R. Flores, M. Lovgren, C. Ardanuy, J. Liñares, D.E. Low, G.J. Tyrrell, J.M. Musser, Full-Genome Dissection of an Epidemic of Severe Invasive

53 ZHAW LSFM, 2017 References Martin Alexander Rappo

Disease Caused by a Hypervirulent, Recently Emerged Clone of Group A Streptococcus, Am. J. Pathol. 180 (2012) 1522–1534. doi:10.1016/j.ajpath.2011.12.037. [65] J. Sjögren, E.F.J. Cosgrave, M. Allhorn, M. Nordgren, S. Björk, F. Olsson, S. Fredriksson, M. Collin, EndoS and EndoS2 hydrolyze Fc-glycans on therapeutic antibodies with different glycoform selectivity and can be used for rapid quantification of high-mannose glycans, Glycobiology. 25 (2015) 1053–1063. doi:10.1093/glycob/cwv047. [66] J. de P. Peter W Atkins, Kurzlehrbuch Physikalische Chemie, 4., WILEY-VCH, Weinheim, 2008. [67] G. Schwedt, Analytische Chemie, 2nd ed., WILEY-VCH, Weinheim, 2008. [68] Y. An, Y. Zhang, H.M. Mueller, M. Shameem, X. Chen, A new tool for monoclonal antibody analysis Application of IdeS proteolysis in IgG domain-specific characterization, MAbs. 6 (2014) 879–893. doi:10.4161/mabs.28762. [69] GE Healthcare Life Sciences, HiScreen TM Capto TM L HiTrap TM Protein L, 1, (2009) 1–20. [70] Genovis AM, FragITkitTM instructions, (n.d.). https://www.genovis.com/wp- content/uploads/instructions-fragit-kit-midi.pdf. [71] Genovis, Immobilized GlycINATOR, (n.d.). https://www.genovis.com/wp- content/uploads/INSTRUCTIONS-for-Immobilized-GlycINATOR®-Midispin.pdf. [72] Merck, Safety data sheet, (n.d.). http://www.sigmaaldrich.com/ (accessed September 19, 2017). [73] U. Kishore, K.B.. Reid, C1q: Structure, function, and receptors, Immunopharmacology. 49 (2000) 159–170. doi:10.1016/S0162-3109(00)80301-X. [74] M.J. Smyth, J.A. Trapani, Granzymes: exogenous porteinases that induce target cell apoptosis, Immunol. Today. 16 (1995) 202–206. doi:10.1016/0167-5699(95)80122-7. [75] R. Rojas, G. Apodaca, Immunoglobulin transport across polarized epithelial cells., Nat. Rev. Mol. Cell Biol. 3 (2002) 944–956. doi:10.1038/nrm972. [76] Merk, Papain, (n.d.). http://www.sigmaaldrich.com/life- science/biochemicals/biochemical-products.html?TablePage=16410606 (accessed September 5, 2017). [77] Sigma-Aldrich, Endoproteinase Lys-C from Lysobacter enzymogenes, (2014) 1–2. https://www.sigmaaldrich.com/content/dam/sigmaaldrich/docs/Sigma/Bulletin/p3428bul.pdf. [78] Genovis, FabULOUS Instructions Version 15.1.2, (n.d.) 1–2. https://www.genovis.com/wp-content/uploads/INSTRUCTIONS-for-FabULOUS- 2000u_version-15.1.2.pdf. [79] Merk, Pepsin from porcine gastric mucosa, (n.d.). http://www.sigmaaldrich.com/catalog/product/sigma/p6887?lang=de®ion=CH&gclid =Cj0KCQjw0ejNBRCYARIsACEBhDNaOVcOkZHowiEafTuS0rmBj8sS9Z8rQeWiRxnh cHt7fFUWnMAfKGUaAmbsEALw_wcB (accessed September 14, 2017). [80] I.M. Rosenberg, Protein analysis and purification: Benchtop techniques, 2., Birkhäuser, Boston MA, 2005. doi:10.1007/b138330. [81] Genovis, GingisREX instructions, (n.d.) 1–2. https://www.genovis.com/wp- content/uploads/INSTRUCTION-gingisrex_17.1.1.pdf. [82] Genovis, Smart Enzymes, (2017). https://www.genovis.com/products/ (accessed September 25, 2017).

54 ZHAW LSFM, 2017 References Martin Alexander Rappo

7.1 List of illustrations

FIGURE 1: AMINO ACID PROPERTIES VENN DIAGRAM [1]...... VIII FIGURE 2: ILLUSTRATION OF THE QUALITY BY DESIGN STRATEGY [15,16]...... 2 FIGURE 3: POTENTIAL INFORMATION SOURCES FOR THE QUALITY RISK MANAGEMENT [15]...... 3 FIGURE 4: THE FOUR QUALITY CATEGORIES FOR THE RISK ASSESSMENT QUALITY ATTRIBUTES [15,17–20]...... 3 FIGURE 5: SELECTED DIGESTION PRODUCTS TO BE EVALUATED DURING THE MASTER THESIS INCLUDING THEIR FINAL PURPOSE IN THE CQA ASSESSMENT...... 4 FIGURE 6: ILLUSTRATION OF THE IGG1 STRUCTURE WITH LIGHT CHAIN IN APRICOT AND HEAVY CHAIN IN BLUE...... 6 FIGURE 7: IG FAB WITH CDR IN RED AND A HAPTEN AS EPITOPE [32]...... 6 FIGURE 8: 3D SURFACE STRUCTURE OF AN IGG1 [33]...... 6 FIGURE 9: NOMENCLATURE OF COMMON IGG FRAGMENTS [37–42]...... 7 FIGURE 10: FC GLYCOSYLATION STRUCTURE OF TRANSGENIC ANIMAL/CHINESE HAMSTER OVARY CELLS...... 8 FIGURE 11: ILLUSTRATION OF IDES...... 9 FIGURE 12: IGG FRAGMENTATION OF IDES...... 9 FIGURE 13: ACCEPTED SEQUENCE FOR THE IDES CLEAVAGE. IDENTICAL AMINO ACID IN BLACK, CLOSELY CHEMICALLY SIMILAR IN BLUE, WEAKLY CHEMICALLY SIMILAR IN SIENNA AND NON- SIMILAR IN APRICOT...... 10 FIGURE 14: PROPOSED CATALYTIC CYCLE OF THE IDES CLEAVAGE...... 10 FIGURE 15: CATALYTIC POCKET AND STABILIZING EFFECTS...... 11 FIGURE 16: ILLUSTRATION OF RECOMBINANT KGP...... 11 FIGURE 17: IGG FRAGMENTATION OF KGP AND IGDE...... 12 FIGURE 18: ACCEPTED KGP CLEAVAGE SEQUENCE OF IGG1 [22]...... 12 FIGURE 19: ILLUSTRATION OF IGDE...... 13

FIGURE 20: ACCEPTED IGDEAGALACTIAE CLEAVAGE SEQUENCE OF HUMAN IGG1 [36]...... 13 FIGURE 21: ILLUSTRATION OF ENDOS2...... 13 FIGURE 22: CHEMICAL STRUCTURE OF THE CHITOBIOSE CORE...... 14 FIGURE 23: CHITOBIOSE CLEAVAGE OF ENDOS2...... 14 FIGURE 24: PROPOSED CONSECUTIVE REACTION SCHEME OF AN IGG AND KGP...... 15 FIGURE 25: SIMULATED KGP DIGESTION PROGRESS BASED ON A CONSECUTIVE FIRST-ORDER KINETIC...... 16 FIGURE 26: DEAMIDATION OF ASPARAGINE...... 18 FIGURE 27: EXAMPLE CATION EXCHANGE CHROMATOGRAM...... 19 FIGURE 28: EXAMPLE CE-SDS NR ELECTROPHEROGRAM OF THE KGP DIGESTION...... 21 FIGURE 29: EXAMPLE CE-SDS RED ELECTROPHEROGRAM OF THE ENDOS2 DIGESTION...... 22 FIGURE 30: FLOW CHART OF THE FAB, FC AND LHF GENERATION AND ISOLATION...... 22 FIGURE 31: FLOW CHART OF THE F(AB)2 GENERATION AND ISOLATION...... 27 FIGURE 32: FLOW CHART OF THE ENDOS2 DIGESTION AND FRAGMENT ISOLATION...... 29 FIGURE 33: PURITY TREND OF THE STABILITY IN DIGESTION BUFFER BY CEX (DEAMIDATION). .... 31 FIGURE 34: PURITY TREND OF THE STABILITY IN DIGESTION BUFFER BY AND CE-SDS NR (FRAGMENTATION/AGGREGATION)...... 31 FIGURE 35: CE-SDS NR ELECTROPHEROGRAM OF THE MIGG1 STABILITY IN DIGESTION BUFFER WITH REDUCING AGENT...... 32 FIGURE 36: CE-SDS NR PURITY PROGRESS OF THE 1:10 KGP-TO-IGG DIGESTION AND COMPARISON OF THE SIMULATED VALUES...... 33 FIGURE 37: REACTION CONSTANT K1 IN DEPENDENCY OF TEMPERATURE FOR TESTED INHIBITORY FACTORS (STANDARD DIGESTION, COMPLETE™ AND PH 5.5)...... 33 FIGURE 38: CALCULATED TIME POINT OF MAXIMAL LHF CONCENTRATION IN DEPENDENCY OF TEMPERATURE FOR TESTED INHIBITORY FACTORS (STANDARD DIGESTION, COMPLETE™ AND PH 5.5)...... 33

55 ZHAW LSFM, 2017 References Martin Alexander Rappo

FIGURE 39: KINETIC CONSTANTS OF THE KGP DIGESTION FOR DIFFERENT CYSTEINE CONCENTRATION OF GENOVIS, SELF-MADE/FRESH AND SELF-MADE/AGED RS...... 34 FIGURE 40: CALCULATED TIME POINT OF MAXIMAL LHF CONCENTRATION OF THE KGP DIGESTION FOR DIFFERENT CYSTEINE CONCENTRATION OF GENOVIS, SELF-MADE/FRESH AND SELF- MADE/AGED RS...... 34 FIGURE 41: PROGRESS OF THE 25 KDA-PEAK FORMATION DURING THE IGG1 KGP DIGESTION AT DIFFERENT CYSTEINE CONCENTRATIONS...... 35 FIGURE 42: KINETIC CONSTANTS OF THE KGP DIGESTION AT DIFFERENT PH...... 36 FIGURE 43: CALCULATED TIME POINT OF MAXIMAL LHF CONCENTRATION OF THE KGP DIGESTION AT DIFFERENT PH...... 36 FIGURE 44: DIGESTION COURSE OF THE PARAMETER ADJUSTED KGP DIGESTION AT PH 6.0...... 37 FIGURE 45: DIGESTION COURSE OF THE PARAMETER ADJUSTED KGP DIGESTION AT PH 8.0...... 37 FIGURE 46: PROTEIN CONTENT OF THE PROTEIN L AFFINITY CHROMATOGRAPHY BY NANODROP. 37 FIGURE 47: CE-SDS NR ELECTROPHEROGRAM OF THE KGP DIGESTION AND CONCENTRATED PROTEIN L FLOW THROUGH...... 38 FIGURE 48: SEC CHROMATOGRAM OVERLAY OF 2 AND 3 COLUMNS IN SERIES...... 39 FIGURE 49: SEC CHROMATOGRAM OF THE INJECTION AMOUNT EVALUATION...... 39 FIGURE 50: SEC CHROMATOGRAM OVERLAY OF THE FRACTION COLLECTION TRIAL OF 20 TIMES 40 ΜG INJECTION...... 40 FIGURE 51: CE-SDS NR OVERLAY OF THE ISOLATED IGG, LHF AND FAB+FC FRACTIONS...... 40 FIGURE 52: SEC CHROMATOGRAM OF THE KGP DIGESTION FRAGMENT ISOLATION WITH AND WITHOUT PROTEIN L AC...... 41 FIGURE 53: SEC CHROMATOGRAM OVERLAY OF THE FRAGMENT ISOLATION AFTER KGP DIGESTION, PROTEIN L AC AND BUFFER EXCHANGE IN MOBILE PHASE...... 42 FIGURE 54: CE-SDS NR ELECTROPHEROGRAM OF INITIAL AND ONE-DAY AGED INJECTION SAMPLE...... 42 FIGURE 55: CE-SDS NR ELECTROPHEROGRAM OVERLAY OF TWO KGP DIGESTIONS WITH IDENTICAL PARAMETER APART FROM DIFFERENT GINGISKHAN® BATCHES...... 43 FIGURE 56: SEC CHROMATOGRAM OVERLAY OF THE FRAGMENT ISOLATION AFTER KGP DIGESTION, PROTEIN L AC AND BUFFER EXCHANGE IN FORMULATION BUFFER...... 43 FIGURE 57: DIGESTION PROGRESS OF IGDE BY CE-SDS NR...... 44 FIGURE 58: ELUTION PROFILE OF PROTEIN L AC AT PH 3.0, 2.5 AND 2.0...... 45 FIGURE 59: CE-SDS NR ELECTROPHEROGRAM OF THE IDES DIGESTION AND ISOLATED PRODUCTS...... 45 FIGURE 60: CE-SDS NR ELECTROPHEROGRAM OVERLAY OF THE THREE FRAGIT™ DIGESTION. 46 FIGURE 61: CE-SDS RED. ELECTROPHEROGRAM OF THE ENDOS2 DIGESTION AND BUFFER EXCHANGE...... 47 FIGURE 62: CEX CHROMATOGRAM OF THE DIGESTION BUFFER STABILITY AT 37°C...... VII FIGURE 63: DIGESTION PROGRESS OF THE DIGESTION ACCORDING TO GENOVIS® STANDARD PROTOCOL...... VII FIGURE 64: KINETIC MODEL PLOT LN([A]/[A]0) IN DEPENDENCY OF TIME TO JUSTIFY A FIRST ORDER KINETIC MODEL...... VIII FIGURE 65: CE-SDS NR ELECTROPHEROGRAM OF THE PROTEIN L AC FLOW THROUGH AND ELUTION...... VIII FIGURE 66: CE-SDS NR ELECTROPHEROGRAM OF THE IDES DIGESTION AND PROTEIN L AC. .... IX FIGURE 67: CE-SDS RED ELECTROPHEROGRAM OF THE DIGESTION BY IMMOBILIZED ENDOS2 AND BUFFER EXCAHNGED SAMPLE...... IX

7.2 List of tables

TABLE 1: ABBREVIATIONS...... VII TABLE 2: LIST OF NEW ICH QUALITY GUIDELINES...... 1 TABLE 3: LIST OF USED BUFFER...... 17 TABLE 4: APPLIED CEX PARAMETER...... 18 TABLE 5: APPLIED SEC PARAMETER FOR FRACTION COLLECTION...... 19 TABLE 6: LIST OF CE-SDS NR PARAMETER...... 20

56 ZHAW LSFM, 2017 References Martin Alexander Rappo

TABLE 7: LIST OF CE-SDS RED PARAMETER...... 21 TABLE 8: DILUTION TABLE OF THE CYSTEINE INFLUENCING DIGESTION...... 24 TABLE 9: ELUTION OF FC ISOLATION...... 25 TABLE 10: SEC FRACTIONATION RANGE...... 25 TABLE 11: ELUTION OF PROTEIN L AC...... 26 TABLE 12: ELUTION OF PROTEIN L AC FOR REPRODUCIBILITY...... 26 TABLE 13: ELUTION OF PROTEIN L AC AFTER IDES DIGESTION...... 28 TABLE 14: PROTEIN L AC PURIFICATION TABLE...... 28 TABLE 15: SUMMARY OF ISOLATED FRAGMENTS DURING SEC TRIALS...... 40 TABLE 16: SUMMARY OF ISOLATED FRAGMENTS DURING SEC INJECTION BUFFER EVALUATION. . 41 TABLE 17: SUMMARY OF ISOLATED FRAGMENTS DURING SEC FRACTION COLLECTION...... 44 TABLE 18: SUMMARY OF ISOLATED FRAGMENTS FOR FRAGIT™ KIT DIGESTION AND FRAGMENT ISOLATION...... 46 TABLE 19: MOST IMPORTANT EFFECTOR FUNCTIONS OF IGG...... II TABLE 20: LIST OF IG FRAGMENTING ENZYMES...... III TABLE 21: LIST OF IG DEGLYCOSYLATING ENZYMES...... III TABLE 22: LIST OF USED INSTRUMENTS...... IV TABLE 23: LIST OF USED CONSUMABLES...... IV TABLE 24: LIST OF USED REAGENTS...... V TABLE 25: LIST OF USED SOFTWARE...... VI

7.3 List of equations

EQUATION 1: DETERMINATION OF THE KINETIC CONSTANT K1 [66]...... 15 EQUATION 2: CALCULATION OF THE INTERMEDIATE PRODUCT CONCENTRATION [I]...... 15 EQUATION 3: CALCULATION OF THE PRODUCT CONCENTRATION [P]...... 15 EQUATION 4: CALCULATION OF THE TIME POINT OF THE MAXIMAL INTERMEDIATE CONCENTRATION IS REACHED...... 16

Specific enzymatic digestion and fragment isolation for the critical quality attribute assessment of IgG1

Annex

Novartis 20.10.2017

Table of content

I. MECHANISM OF ACTION ...... II II. ENZYMES IN MAB ANALYTICS ...... III I.1. FRAGMENTATION ...... III I.2. DEGLYCOSYLATION ...... III III. INSTRUMENTS ...... IV IV. CONSUMABLES ...... IV V. REAGENTS ...... V VI. ENZYMES ...... VI VII. SOFTWARE ...... VI VIII. CHROMATOGRAM AND ELECTROPHEROGRAM ...... VII I.3. FAB, FC AND LHF ...... VII I.4. F(AB’)2 ...... IX I.5. CHITOBIOSE CLEAVAGE ...... IX IX. COPY OF THE OBJECTIVE SETTING AGREEMENT ...... X X. STATEMENT OF AUTHORSHIP ...... XII XI. REPORT IN DIGITAL FORMAT ...... XIII XII. POSTER ...... XIV

I ZHAW LSFM, 2017 Annex Martin Alexander Rappo

I. Mechanism of action

Table 19: Most important effector functions of IgG.

Function Mechanism Complement activation After epitope binding on a cell an IgG undergoes a conformational change. Two bound Igs next to each other allow for binding of 318 320 322 C1q, a protein complex, at the CH2 (Glu , Lys , Lys ). Subsequently, the complement cascade is initiated, which leads to pore formation and finally the destruction of the cell [25,73]. Ab-dependent cellular Cell bound IgGs allow natural killer-cells to bind the Fc region with cytotoxicity (ADCC) a Fc receptor (FcR), the FcRs are cross-linked and initiate the release of granzymes and leading to apoptosis of the cell [25,74]. Opsonization Igs bound to bacteria present their Fc part to phagocytes and simplify the internalization [25]. Blocking Ig binding to a receptor may prevent the access for ligands and their physiological function [25]. A special form of blocking is the binding of an Ig to the CDR of another Ig, so called anti idiotypic Ab. This leads to the elimination of this respective Ab [25]. Masking Binding of non-cytotoxic Ig may prevent cytotoxic T-cells resp. Igs to recognize and bind an antigen [25]. Neutralization Small molecular toxins might act as an antigen and through binding by an Ig the binding to a receptor and the toxic effect is prevented [25]. Receptor binding Ab binding to a receptor can have three possible effects: No reaction The receptor is not effected by the Ab [25]. Agonistic Ab The Ig binding itself imitates a ligand binding and triggers a response. This effect is only relevant for auto-immune disease [25]. Antagonistic Ab The Ig blocks the access of the receptor for ligands and prevents a physiological reaction [25]. Inhibition Some specific soluble immune complexes have the ability to suppress the activation of B-cells. This is important to reduce the antigen specific B-cell response [25]. Penetration Igs are able to cross polarized epithelial cells through cell- mediated transcytosis. For IgG the neonatal Fc receptor (FcRn) is of importance and even allows passing of the placenta [25,75]. Precipitation Binding of polyclonal Ab on multiple epitopes first forms soluble complexes and leads later to precipitation [25]. Agglutination Linking multiple particular antigens, e.g. on the surface of cells, leads to clump [25].

II ZHAW LSFM, 2017 Annex Martin Alexander Rappo

II. Enzymes in mab analytics

I.1. Fragmentation

Table 20: List of Ig fragmenting enzymes.

Application Name Trade name Information Fab + Fc Papain Papain Cleaves hydrophilic-R/K ↓ not V [76]. Over-digestion possible. Lys-C Lys-C Cleaves C-terminal of K. Over- digestion possible. [77] SpeB FabULOUS™ Cleaves human IgG1: KTHT ↓ CPPCPAPEL under reducing conditions. [78] Kgp GingisKHAN® Cleaves human IgG1: KSCDK ↓ THTCPPCP under mild reducing conditions. IgdE FabALACTICA™ Cleaves IgG1: KSCDKT ↓ HTCPPCP [23]

F(ab’)2 Pepsin Pepsin Cleaves the peptide bond of hydrophobic and aromatic amino acids [79]. Risk of oxidation. IdeS FabRICATOR® Cleaves human IgG1, 3 and 4: CPAPELLG ↓ GPSVF and human IgG2: CPAPPVA ↓ GPSVF. [21] Peptide mapping Trypsin Trypsin Cleaves C-terminal of K/R-not P [80]. Rgp GingisREX™ Cleaves C-terminal of R [81]. Removal of C- Carboxypeptidase Carboxypeptidase Cleaves K and R at the C-terminus terminal K B B [80].

I.2. Deglycosylation

Table 21: List of Ig deglycosylating enzymes.

Cleavage site Name Trade name Information -GlcNAc | Asn- PNGase F PNGase F Complex, high-mannose and hybrid type oligosaccharides of N-linked glycoproteins [80]. -GlcNAc | GlcNAc- EndoS IgGZERO® Only complex oligosaccharides of IgG and the α1-acid glycoprotein [82]. -GlcNAc | GlcNAc- EndoS2 GlycINATOR® Complex, high-mannose and hybrid type oligosaccharides of IgG and the α1-acid glycoprotein [82].

III ZHAW LSFM, 2017 Annex Martin Alexander Rappo

Removal of α2,3-, Two sialidase of SialEXO™ Cleaves N- and O-glycans[82]. α2,6- and α2,8- Akkermansia linked sialic acids muciniphila O-glycosylated endo-α-N- OglyZOR™ Required prior sialic aids proteins of core 1 acetylgalactos- removal [82]. and core 3 aminidase O-glycosylated Endoprotease of OpeRATOR™ Improved enzyme activity with proteins N- Akkermansia removed sialic aids [82]. terminally of the muciniphila S/T glycosylation site

III. Instruments

Table 22: List of used instruments.

Type Model Manufacturer Centrifuge 0.5 mL tubes 1-14 SIGMA Centrifuge 4 & 15 mL tubes 3-18KS SIGMA CE-SDS NR/Red. PA800plus Pharmaceutical BECKMAN COULTER Analysis System HPLC (CEX and SEC) 1100 Series Agilent Technology HPLC fraction collector 1260 Infinity Agilent Technology Thermo shaker Thermomixer comfort EPPENDORF Tube rotator plate Tube rotator BOEKEL Water bath Eco Silver Lauda Nanodrop™ NanoDrop1000 Thermo Scientific™ Spectrophotometer

IV. Consumables

Table 23: List of used consumables.

Description Manufacturer Item number Amicon® Ultra – 0.5mL Merk UFC501096 Centrifugal Filters Ultracel® - 10K Amicon ® Ultra – 4 Merk UFC801024 Centrifugal Filters Ultracel® - 10K Amicon ® Ultra – 15 Merk UFC901096 Centrifugal Filters Ultracel® - 10K Stericup® 150ml Millipore SCGPU01RE Millipore® Express PLUS

IV ZHAW LSFM, 2017 Annex Martin Alexander Rappo

0.22µm PES Protein L affinity GE Healthcare 29048665 chromatography HiScreen™ Capto™ L HiTrap™ Protein L, 1 mL Crimp vials 6 mL + 20 mm Agilent Technology 9301-1419 + 9301-1425 crimp caps with septum

V. Reagents

Table 24: List of used reagents.

Name Supplier Part number Purity / Conc. MW in g/mol 2- SIGMA-ALDRICH 63689 ≥ 99.0% 78.13 mercaptoethanol cOmplete™ Roche 11 697 498 001 n.a. n.a. protease inhibitor cocktail Cysteine SIGMA-ALDRICH 30089 98.5% 121.16 Hydrochloric acid Merk 1.09060 1 M 36.46 Iodoacetamide SIGMA-ALDRICH I1149 ≥ 99% 184.96 L-Histidine SIGMA-ALDRICH H8000 ≥ 99% 155.15 L-Histidine SIGMA-ALDRICH H8125 ≥ 98% 209.63 hydrochloride monohydride ortho-phosphoric Merk 1.00552 85% 97.99 acid Potassium SIGMA-ALDRICH P9541 ≥ 99.0% 74.55 chloride SDS-MW Gel BECKMAN A30341 n.a. n.a. buffer COULTER Sodium chloride Honeywell | Fluka 71380 ≥ 99.5% 58.44 Sodium hydroxide Merk 1.09137 1 M 40.00 solution Sodium SIGMA-ALDRICH S7907 ≥ 99.0% 141.96 phosphate dibasic anhydrous Sodium SIGMA-ALDRICH 71507 ≥ 99.5 137.99 phosphate monobasic monohydrate Trifluoroacetic Thermo Scientific 28903 Sequencing grade 114.02 acid Tris-HCl Promega H5123 ≥ 99.0% 157.56

V ZHAW LSFM, 2017 Annex Martin Alexander Rappo

Trisodium citrate SIGMA-ALDRICH S1804 100% 294.10 dihydrate Trizma® base SIGMA-ALDRICH T1503 99.9% 121.14 UltraPure™ Tris- Gibco 15568025 1 M 157.60 HCl pH 8.0

VI. Enzymes

Name Supplier Part number Amount in units MW in kDa FabALACTICA™ Genovis A0-AG1-020 2000 70 FabRICATOR® Genovis A0-FR1-020 2000 38 FragIT™ kit Genovis A0-FR6-100 n.a. 38 midispin GingisKHAN® Genovis B0-GKH-020 2000 50 GlyCINATOR® Genovis A0-GL1-020 2000 92 Immobilized Genovis A0-GL6-100 n.a. 92 GlycINATOR® MidiSpin

VII. Software

Table 25: List of used software.

Name Version Developer ChemDraw Ultra 14.0 PerkinElmer Chromeleon 6.8 Thermo Scientific™ GPMAW 10.0 Lighthouse data Open Lab CDS C.01.05 [41] Agilent Technologies Chemstation Edition for LC & LC/MS Systems PyMOL Molecular Graphics 1.5.0.5 SCHRÖDINGER System

VI ZHAW LSFM, 2017 Annex Martin Alexander Rappo

VIII. Chromatogram and electropherogram

I.3. Fab, Fc and LHF

Stability evaluation in digestion buffer with and without reducing solution

CEX chromatogram - Digestion buffer stability at 37°C

T0@37°C T1h@37°C T2h@37°C T24h@37°C Absorbance in mAU

12 13 14 15 16 17 18 19 20 21 22 Time in min.

Figure 62: CEX chromatogram of the digestion buffer stability at 37°C.

Genovis protocol digestion procedure 100%

90%

80%

70%

60% IgG 50% LHF 40% Purity Purity in A% Fc

30% Fab

20%

10%

0% 0 15 30 45 60 75 90 105 120 Time in min

Figure 63: Digestion progress of the digestion according to Genovis® standard protocol.

VII ZHAW LSFM, 2017 Annex Martin Alexander Rappo

Kinetic plot: ln([A]/[A]0) VS. time 0.5

0 10 20 30 40 50 60 -0.5

-1.5

y = -0.0838x - 0.2684 0 R² = 0.9683 -2.5 ln([A]/[A]0)

Linear (ln([A]/[A]0)) ln([A]/[A] -3.5

-4.5

-5.5 Time in min.

Figure 64: Kinetic model plot ln([A]/[A]0) in dependency of time to justify a first order kinetic model.

CE-SDS NR: Pooled and conentrated Protein L AC fractions

Protein L Wash Protein L elution Absorbance in mAU

9 10 11 12 13 14 15 16 17 Time in min.

Figure 65: CE-SDS NR electropherogram of the Protein L AC flow through and elution.

VIII ZHAW LSFM, 2017 Annex Martin Alexander Rappo

I.4. F(ab’)2

CE-SDS NR electropherogram of the Protein L affinity chromatography and F(ab')2 isolation

Reference 1h digestion Flow through

Wash Elute pH 2.5 Elute pH 2.0 Absorbance in mAU

7 8 9 10 11 12 13 14 15 16 17 Time in min.

Figure 66: CE-SDS NR electropherogram of the IdeS digestion and Protein L AC. I.5. Chitobiose cleavage

CE-SDS red. electropherogram of the immob. EndoS2 chitobiose cleavage and buffer exchange

Reference EndoS2 digestion Buffer exchange Absorbance in mAU

15 20 25 30 35 40 45 50 Time in min.

Figure 67: CE-SDS red electropherogram of the digestion by immobilized EndoS2 and buffer excahnged sample.

IX ZHAW LSFM, 2017 Annex Martin Alexander Rappo

IX. Copy of the objective setting agreement

X ZHAW LSFM, 2017 Annex Martin Alexander Rappo

XI ZHAW LSFM, 2017 Annex Martin Alexander Rappo

X. Statement of authorship

I certify hereby, that I have written this report autonomous, without outside assistance and I have used only allowed resources. I affirm particularly, that I have mentioned all literal and paraphrase transfers as such and assigned the original author and document.

Any violation will lead to actions according to §39 and §40 of the General Academic Regulations for Bachelor’s and Master’s degree programs at the Zurich University of Applied Sciences of 29.01.2008.

Location

Date Signature

XII ZHAW LSFM, 2017 Annex Martin Alexander Rappo

XI. Report in digital format

XIII ZHAW LSFM, 2017 Annex Martin Alexander Rappo

XII. Poster

XIV