Analytical Characterization of Biotherapeutics ebook

Table of Contents

Cover

Title Page

List of Contributors

1 Introduction to Biotherapeutics

1.1 Introduction

1.2 Types of Biotherapeutics and Manufacturing Systems

1.3 Types of Analyses Performed

1.4 Future perspectives

Acknowledgments

References

2 Mass Spectrometric Characterization of Recombinant Proteins

2.1 Introduction

2.2 Peptide Mass Fingerprinting

2.3 Tandem Mass Spectrometric Characterization of Biomolecules

2.4 Conclusions and Perspectives

References

3 Characterizing the Termini of Recombinant Proteins

3.1 Introduction

3.2 Gel Electrophoresis and Edman Sequencing

3.3 Mass Spectrometric Approaches for Characterizing True Starts of Proteins

3.4 Conclusions

References

4 Assessing Activity and Conformation of Recombinant Proteins

4.1 Introduction

4.2 Circular Dichroism

4.3 DSC and Isothermal Titration Calorimetry

4.4 Hydrogen–Deuterium Exchange–Mass Spectrometry

4.5 Nuclear Magnetic Resonance

4.6 Concluding Remarks

References

5 Structural Characterization of Recombinant Proteins and Antibodies

5.1 Introduction

5.2 Antigens, Epitopes, and Paratopes

5.3 Choice of Analytical Method for Epitope Mapping

5.4 Recombinant Antigen Generation

5.5 N‐linked Glycosylation

5.6 Antibody Generation for Crystallography

5.7 Crystallization of Antibody/Antigen Complexes

5.8 Conclusion

References

6 Antibody

de novo

Sequencing

6.1 Introduction

6.2 Technical Details on Antibody

de novo

Sequencing

6.3 Bioinformatics Workflow

6.4 Sequence Validation

6.5 Conclusions

References

7 Characterization of Antibody–Drug Conjugates

7.1 Introduction

7.2 Characterization of DAR Utilizing MS

7.3 Structural Characterization of ADC

7.4 Characterization of ADC Catabolism by MS

7.5 Conclusions

References

8 Characterization of Bispecific or Other Hybrid Molecules

8.1 Introduction

8.2 Overview of the Various Bispecific Formats

8.3 Alternatives to Bispecific Antibodies: Antibody Mixtures

8.4 Characterization of the Bispecific Molecule

8.5 Conclusions

References

9 Bio‐Repository

9.1 Introduction

9.2 Large Molecule Repository Management

9.3 Challenges and Future Perspectives for Working with Diverse Biological Reagent Types

References

10 Characterization of Residual Host Cell Protein Impurities in Biotherapeutics

10.1 Introduction

10.2 HCP Measurement and Reporting

10.3 Methods to Characterize Host Cell Impurities

10.4 Use of HCP‐ELISA and Orthogonal 1D LC‐MS/MS in Practice

10.5 Risk of HCPs Present in Products

10.6 Conclusions

References

11 Analytical Tools for Biologics Molecular Assessment

11.1 Introduction to Molecular Assessment

11.2 Molecular Assessment

11.3 Biotherapeutic Stability

11.4 Physical Degradation

11.5 Yield and Structural Stability

11.6 Posttranslational Modifications

11.7 Analytical Techniques

11.8 Summary

References

12 Glycan Characterization

12.1 Introduction

12.2 Glycan Labeling

12.3 Compositional Analysis

12.4 Glycan Release

12.5 Determining Sites of Glycosylation

12.6 Determining N‐linked Glycan Distribution

12.7 Comparison of Methods Used in Determining Glycan Distribution

12.8 Assessing N‐linked Glycan Structure

References

Index

End User License Agreement

List of Tables

Chapter 01

Table 1.1 The circulatory half‐life of a therapeutic protein can be extended by several strategies depending upon the endogenous clearance mechanism of the drug.

Chapter 08

Table 8.1 Production technologies of bispecific antibody formats and their advantages and limitations.

Table 8.2 Techniques to purify and characterize bispecific antibodies.

Chapter 11

Table 11.1 Typical molecular assessment activities.

Chapter 12

Table 12.1 Glycans observed in a recombinant mAb produced in CHO cells.

Table 12.2 Comparison of common methods used in determining glycan distribution.

List of Illustrations

Chapter 01

Figure 1.1 Various categories of the main types of biotherapeutics currently marketed.

Figure 1.2 Monoclonal antibody (mAb) structure can be modified on the basis of the desired mechanism of action. Immunoglobulin G1 (IgG1) is the most effective naturally occurring human IgG isotype at mediating antibody‐dependent cell‐mediated cytotoxicity (ADCC). Glycomodified afucosylated mAbs (part a) (such as Obinutuzumab) demonstrate enhanced binding to IgG Fc receptors (FcγRs) and enhanced ADCC. In addition antibody‐dependent cellular phagocytosis, a process mediated by macrophages, can also occur [8]. Afucosylated mAbs are produced using cell lines that lack the enzymes responsible for fucosylation. Modifying the amino acid sequence of mAb Fc (part b), as was done to produce ocaratuzumab [9], can also result in enhanced binding to FcγRs and enhanced ADCC. For mechanisms of action in which ADCC is not desirable, IgG4 may be a more appropriate isotype, as IgG4 mAbs do not mediate ADCC to the same degree as IgG1 (part c) although this isotype can still engage macrophage effector function

via

nanomolar affinity binding to FcγRI. Nivolumab, an IgG4 mAb that blocks programmed cell death protein 1 (PD1) on T cells, is one such example. Producing radioimmunoconjugates involves linking the radioisotope to the mAb. A stable linker is most desirable (part d) to limit the leakage of the free radioactive isotope. Conversely, optimal antibody–drug conjugates (ADCs) use a cleavable linker (part e). To avoid nonspecific toxicity, it is desirable for drugs used in ADCs to be cytotoxic once inside the target cell but nontoxic when bound to the mAb in the circulation. Linkers that are pH‐sensitive or enzymatically cleaved are now a standard component of ADCs. Chimeric antigen receptor (CAR) T cells get their specificity from mAb variable regions but are a form of gene, not protein, therapy. They are produced by inserting DNA coding for the mAb variable region fused to DNA coding for signaling peptides into T cells (part f). Some bispecific antibodies lack a functional constant region so that they do not nonspecifically crosslink activating receptors and activate T cells (part g). The lack of a constant region on such constructs results in a short half‐life, thus requiring continuous infusion to achieve the desired exposure.

Chapter 02

Figure 2.1

MALDI ionization

. The uncharged analyte is mixed with an uncharged matrix molecule. After receiving energy

via

the laser pulse the now charged matrix donates a proton to the analyte

via

desorption, therefore allowing the analyte to become charged. The charged analyte is guided through the MALDI mass spectrometer for

m

/

z

analysis.

Figure 2.2

High‐mass MALDI‐TOF spectrum of IgA dimer

. Secretory immunoglobulin A (sIgA) exists as a dimer and is the main immunoglobulin found in mucus secretion. A high‐mass linear detector (CovalX, Saugus, MA) attached to a 4800 MALDI‐TOFTOF (Sciex, Redwood City, CA) enabled the detection of a 362 kDa glycosylated IgA dimer and demonstrated that the ratio of antibody to J‐chain was 2 : 1. Doubly (181 kDa) and triply (120 kDa) charged species are also observed.

Figure 2.3

Schematic of electrospray ionization

. (1) Under high voltage, the eluent from a syringe, tip, or HPLC creates a cone shape (a Taylor cone) and represents the initiation of the transfer of ions from the liquid to the gas phase. (2) Evaporation of the droplets occurs leaving them increasingly highly charged. (3) When the charge exceeds the Rayleigh limit (the maximum amount of charge a liquid droplet can carry before throwing out fine jets of liquid), the droplet completely dissociates leaving a stream of charged gaseous ions that now enter the mass spectrometer for separation and detection.

Figure 2.4

Schematic of a quadrupole time of flight mass spectrometer

. An analyte, for example, a protein of interest undergoes ionization at the ion source and gaseous ions are transmitted into the mass spectrometer. After ions enter the source, the pressure is raised to increase transmission efficiency. Ions then pass through the quadrupole (where mass selection can take place) before being accelerated into the collision cell. The collision cell is filled with an inert gas. Analytes are separated through the flight tube where they are resolved based upon the time it takes them to traverse across a predefined distance. To enable higher resolution using reflectron mode, ions are “reflected” back along the flight tube and detected by the second detector.

Figure 2.5

Native MS analysis of and RGY–antibody hexamer

. Under native conditions (10 mM ammonium acetate) on an EMR Exactive Orbitrap, using high C‐trap pressure the RGY‐IgG mutant is observed to be a hexameric species. The charge envelope of the IgG monomer appears around

m

/

z

6000 and that of the IgG hexamer around

m

/

z

13 000.

Figure 2.6

Schematic of a conventional drift time IMS

(

DTIMS

) system showing three ions of different sizes in the reaction region and then migrating at different velocities in the drift region.

Figure 2.7 Schematic of a high‐performance liquid chromatography (HPLC) system: Analytes are separated on the solid stationary phase by displacement with a liquid mobile phase. The stationary phase is the packing material contained in the HPLC column (hollow tube containing particles). The mobile phase is the solvent or mixture of solvents (and sometimes other ingredients) that is passed through the HPLC column to flush the compounds through the column. Different compounds have different affinities that are attracted to the particles in the HPLC column and different affinities to being flushed through the column using the mobile phase. The competition between a compound’s affinity to bind to the stationary phase or elute away in the liquid mobile phase is what enables the separation. A mixture of different compounds can be separated by chromatography and then detected individually by a detector (e.g., UV absorbance). The absorbance response for the separated compound will be proportional to the concentration of the compound in solution. Comparison of the detector response for an unknown to that of a known standard (known concentration of a pure compound) can determine the concentration present in the unknown. The analytes (compound you are testing the sample for) of interest must be dissolved in a solvent to perform an HPLC separation.

Figure 2.8

Schematic of a capillary electrophoretic system

. The analyte of interest is introduced into the CE system by placing the capillary inlet into the samples vial where it traverses through the capillary by capillary action, pressurization, siphoning, or through electrokinetic mechanisms, where finally the sample exits at the source vial. Migration of analytes occurs when an electric field (supplied by the electrodes) is applied between the source and destination vials and in most CE systems all ions, positive or negative, navigate through the capillary in the same general direction by electroosmotic flow. Analytes are detected near the outlet end of the capillary and data is displayed as an electropherogram.

Figure 2.9 Agilent microfluidic device.

Figure 2.10

Analysis of deglycosylated antibody with on‐chip deglycosylation

. Three main peaks are observed prior to deglycosylation. These peaks are suggestive of the antibodies with the combination of G0, G1, and G2 glycans. The partially deglycosylated antibody follows a 3 s residence time. Antibodies with one glycan are observed. The deglycosylated antibody peak after a 6 s residence time in the PNGase F enzyme reactor chip.

Figure 2.11

Schematic of a peptide mass fingerprinting workflow

. The sequence of a protein of interest is selected from a database. An

in silico

digestion is performed (e.g., in this case trypsin is employed, assuming cleavage at the C‐terminus of Arginine and Lysine). Theoretical masses are then assigned to these peptides and, either manually or using an algorithm, predicted masses are correlated with experimental masses, allowing researchers to determine if the protein being analyzed matches the protein in the database.

Figure 2.12

Overview of top‐down and bottom‐up MS

‐based protein biophysical studies (using antibody as example). The left circle is the summary of top‐down approaches. The right circle is the summary of bottom‐up approaches.

Figure 2.13

Peptide fragment MS

/

MS ion series

. If charge is retained on the N‐terminal fragment, the ion is classed as

a

,

b

, or

c

. If the charge is retained on the C‐terminal, the ion type is

x

,

y

, or

z

. A subscript indicates the number of residues in the fragment.

Chapter 03

Figure 3.1 Edman sequencing. Proteins and peptides (shown with each separate amino acid residue as circle) with freely available N‐termini can be reacted with phenylisothiocyanate (PITC) (marked as a triangle) under alkaline conditions. This generates a PITC–peptide derivative which after rearrangement and treatment with acid and heat results in the peptide (now one residue shorter) and a thiozalinone group (specific to the first residue of the original peptide). This group can be extracted in organic solvent and rearranged into a phenylthiohydantoin (PTH) group (triangle/circle) with acid and heat treatment. The residue‐specific PTH group can be identified by chromatography or electrophoresis and the isolated peptide is available for the subsequent cycle of sequencing as indicated.

Figure 3.2 Top‐down proteomics. Intact protein mass spectrometry (left hand side) is contrasted to peptide mass spectrometry (on the right hand side). Biological molecules are directly infused to the mass spectrometer (shown here by ESI). Protein ions often exist in multiple different charge states (seen here between +5 and +10), thus increasing mass spectra complexity, whereas peptides often exist in one or two charge states only (+2 and +3 shown here). In addition, when zoomed in it can be seen that the +3 ion shows a simple isotopic distribution where the monoisotopic peak (marked by an arrow) is easily determined for correct mass‐to‐charge calculations. Contrastingly, the isotopic pattern for the +10 ion is far more complex and the monoisotopic peak (marked by an arrow) is not as easily determined due to the requirements for higher‐resolution instrumentation to differentiate across all peaks and high sensitivity to determine the monoisotopic peak with a confident signal‐to‐noise ratio. Each isolated precursor ion can be fragmented with a variety of methods. For top‐down mass spectrometry, the exact mass of the ion is used (hence the critical need for correct monoisotopic peak determination) in combination with sequence information generated by fragmentation events, which can be used to interrogate databases for the N‐terminal sequence of the protein.

Figure 3.3 ATOMS. (a) ATOMS can be utilized to identify the true N‐terminus of a recombinant protein by labeling the entire protein by reductive dimethylation. Retaining positional information by means of the tag allows digestion with trypsin (and a separate aliquot with lysargiNase) to generate peptides that can be analyzed by LC‐MS/MS. Peptides with a dimethylated N‐terminus are indicative of the true start of the protein. (b) A protein N‐terminus or a protease‐cleaved recombinant protein substrate of interest can be analyzed by ATOMS as shown on the right hand side. Primary amines are blocked at the protein level with different combinations of isotopically labeled formaldehyde and cyanoborohydride (shown as circles). In this example the newly generated N‐terminus from protease cleavage is indicated in dark gray. Each labeled protein sample can then be mixed and digested with trypsin (and a separate aliquot with lysargiNase) to generate peptides, which as highlighted are a combination of freely available primary amines or blocked by dimethylation. Note that natural N‐termini are blocked at equal ratios of dimethyl tags whereas the newly formed N‐terminus from protease cleavage is a singlet. Following LC‐MS/MS analysis it can be seen that nonlabeled tryptic peptides cannot be quantified, the protease‐cleaved N‐terminal peptide exists in a heavy: light ratio ≫ 1, while natural protein N‐termini are dimethylated at their peptide N‐termini in a ratio of 1 : 1. Internal tryptic peptides that have been labeled also are present in a ratio of 1 : 1. Database searching and quantification allows the identification of the protease‐cleaved N‐terminus as well as the natural N‐termini. However, a simple rule is that any peptide with a labeled N‐terminal primary amine represents the N‐terminus or Neo‐termini present in the sample.

Figure 3.4 LysargiNase digestion to profile true C‐termini of recombinant proteins. (a) Digestion of a recombinant protein with trypsin yields fully specific internal peptides (protease cleavage site at the N‐terminus and C‐terminus of the peptide) and semi‐specific terminal peptides (either the N‐terminus or the C‐terminus have the protease cleavage site). Most proteins do not contain a positively charged C‐terminal residue, thus making tryptic‐generated C‐terminal peptides unfavorable for mass spectrometric analysis. (b) Digestion of the same recombinant protein with lysargiNase generates N‐terminal basic residues that are amenable for mass spectrometric analysis. In this case the C‐terminus of the protein has a peptide that can be observable by MS analysis and provides the true end of the protein.

Chapter 04

Figure 4.1 Hierarchy of biophysical methods for the characterization of higher order structure of proteins. The methods are organized in four tiers according to the degree of complexity in the experimental setup and level of resolution they provide. AUC, analytical ultracentrifugation; CD, circular dichroism; DLS, dynamic light scattering; DSC, differential scanning calorimetry; FTIR, Fourier transform infrared; HDX‐MS, hydrogen–deuterium exchange–mass spectrometry; NMR, nuclear magnetic resonance; SAXS, small angle X‐ray scattering; SEC, size‐exclusion chromatography; UV, ultraviolet.

Figure 4.2 Far UV CD spectra associated with various types of secondary structure. Solid line, α‐helix; long dashed line, antiparallel β‐sheet; dotted line, type I β‐turn; cross dashed line, extended 3

1

helix or poly (Pro) II helix; short dashed line, irregular structure.

Figure 4.3 Thermal stability analysis of WT (left panel) and engineered Fc fragments (center and right panels). The lower

T

m

corresponds to the melting of the C

H

2 domain and the higher

T

m

corresponds to the C

H

3 domain.

Figure 4.4 Analysis of thermal stability of antibodies by DSF compared to DSC. (a) Thermogram of an antibody obtained by DSF. The different

T

m

correspond to the inflection points of the curves, labeled as

T

m

1 and

T

m

2. (b) The different transition temperatures are identified more easily as the peaks of the first derivative function of the thermogram. (c) DSC analysis of the same antibody. (d) Correlation between the

T

m

obtained by DSC or DSF for three antibodies in different formulations.

Figure 4.5 Applications of HDX‐MS in biopharmaceutical drug development. Left panel: Mapping of protein–drug and protein–protein interaction sites and related conformational changes. Middle panel: Monitoring the conformational response of protein pharmaceutical to different formulations. Right panel: Comparing the structure of therapeutics produced by different processes.

Figure 4.6 Epitope mapping by HDX. The inset show six different overlapping peptides that show reduced HDX over time in the presence of mAb 12C1. The segments of the protein corresponding to the protected peptides are highlighted in red.

Figure 4.7 Process for determining the 3D structure in a NMR experiment. Experimental parameters are fed into a computer and run through a variety of optimizing programs to generate an ensemble of structures that satisfy the experimental constrains.

Figure 4.8 2D

1

H–

13

C methyl SOFAST‐HMQC NMR spectra for three different commercial insulin products. Top panel: overlay of Umuline® and Actrapid® showing very similar profile. Bottom panel: Overlay of Umuline and Humalog® spectra with several chemical shift changes indicated with arrows.

Chapter 05

Figure 5.1 Perjeta (red/yellow) and Herceptin (blue/green) bound to the EGF family receptor HER2 (overlaid and colored in purple), illustrating the unique blocking epitopes bound by each of these marketed antibodies. PDB IDs are 1S78 [3] and 1NZ8 [2].

Figure 5.2 A stalk epitope on hemagglutinin (HA) that neutralizes most strains of influenza A. (a) Consurf analysis showing the conservation of surface residues in an H3 HA, displayed on a scale from purple (most conserved) to teal (least conserved). Heavy (yellow) and light (green) chains of the 39.29 antibody [5] are displayed to show the position of the epitope. (b–d) Epitopes of broadly neutralizing antibodies 39.29 bound to H3 HA [5] (b), FI6v3 bound to H1 HA [6] (c) and CR9114 bound to H5 HA [7] (d). Heavy chain epitope is shown as yellow surface and light chain epitope as green surface.

Figure 5.3 Comparison of antibody epitopes on VEGF. Surface representation shown for (a) VEGFR1 (magenta) bound to VEGF (orange) [92], (b) the G6 Fab (blue) bound to VEGF [9], (c) the B20‐4 Fab (green) bound to VEGF [9], and (d) Avastin Fab bound to VEGF [10].

Figure 5.4 2D [

1

H–

15

N]–HSQC spectrum of uniformly

15

N–

13

C‐labeled and 50% deuterated ubiquitin. Peaks are labeled according to the corresponding amino acid number in the protein sequence. Resonances belonging to the asparagine and glutamine side chain amide protons are connected by dashed lines.

Figure 5.5 Epitope mapping by NMR. (a) Superposition of 2D

1

H–

15

N–HSQC spectra of free (red) and Fv‐bound

15

N–

13

C–

2

H‐labeled EZ4, a mutant of the

Staphylococcal

protein A (SpA) domain E, in the free (red) and Fv‐bound (blue) forms. Peaks undergoing a large change in chemical shifts are labeled. (b) Residues that experience significant chemical shift changes upon complex formation are colored in blue on the structure of the E‐domain.

Chapter 06

Figure 6.1 Zoomed‐in views of the C‐terminal end of the primary heavy chain variable region of a hybridoma sequenced with next generation sequencing (NGS) (a) and MS/MS (b) technologies. The sequence is shown at the top of each image. Each blue line denotes a peptide identified by a database search tool (MSGFDB) [2] (at <0.1% false discovery rate (FDR)) against three MS/MS spectra from the same precursor (CID, HCD, and ETD). Without MS/MS sequencing efforts, two sequencing errors in the framework 4 (FR4) region would have gone unnoticed: GQGTLVTVSSASTK → GQGTMVTVPSASTK. These figures were generated with the pViz software [3].

Figure 6.2 (a) Complementary fragmentation in CID and ETD for peptide TAAANAAAGAAENAFRAP. CID and ETD spectra were separately identified against this C‐terminal tryptic peptide at 1% FDR. Enough ions were separately detected in each spectrum to identify the peptide (65% of breaks in CID, 53% in ETD). But combining the two yields full coverage of all possible breaks, thus giving higher confidence to breaks observed in both spectra and possibly enabling full‐length

de novo

sequencing. (b) (Left) Peptide MS2 ion statistics for alternative fragmentation modes—This shows the percentage of breaks observed by each ion type over all identified MS2 spectra with precursor charge 2 or 3 for each fragmentation method. z° corresponds to peaks at offset +H from z ions. Ions were counted if observed peak masses were within 20 ppm of expected ion masses. The “noise” ion corresponded to offset b + 0.5, which was counted to show the level of noise in each type of MS2 spectra. (Right) Peptide break statistics for combinations of alternative fragmentation modes—peptide breaks were counted for all unique peptides identified by all three fragmentation modes. The six columns show the percentage of breaks detected by each fragmentation mode and combination of fragmentation modes per precursor charge state. In CID and HCD spectra, the presence of breaks was indicated by the presence of b‐ or y‐ions. For ETD, c, z°, or z°+H ions indicated the presence of a break. Multiply charged ions (up to the spectrum’s precursor charge) were also considered in each spectrum. Prior to this analysis, peak filtering was applied to all CID, HCD, and ETD spectra such that each peak was retained only if its intensity was ranked fifth or higher over all neighboring peaks in a ±56 Da radius. If a peptide was identified by more than one CID, HCD, or ETD spectrum, a single representative spectrum was randomly chosen for each fragmentation mode.

Figure 6.3 Discovery and identification of posttranslational modifications through spectral networks; (a) spectral alignment between modified and unmodified variants of the peptide TETMA (

b

‐ions shown in light gray,

y

‐ions in dark gray, black lines track consecutively matched

b

/

y

‐ions); (b) grouped modification states of the peptide MDVTIQHPWFK from a sample of cataractous lenses. Nodes in the spectral network represent individual MS

2

spectra and edges between nodes represent significant spectral alignments such as that shown in part (a); (c) spectra assembled in the spectral network, or

contig

, for TNSMVTLGCLVK with diverse cysteine modifications on a monoclonal antibody. Each arrow corresponds to the propagation of a sequence and/or PTM from an identified spectrum to an unidentified spectrum (repeated arrows are iterative propagations). Shaded arrows correspond to types of modifications transferred.

Chapter 07

Figure 7.1 Dotmatics Vortex processing of the deconvoluted intact protein mass spectra of a mixture of ADC and its catabolites. Features were added to the software to reconstruct the deconvoluted mass spectra and automatically label the observed antibody, linker–drug moieties, as well as the possible modifications. Antibody graphics have been added manually for illustrative purposes.

Chapter 08

Figure 8.1 Bispecific antibody modes of action. (a) Retarget cytotoxic T‐cells or NK cells to kill tumor cells, (b) inactivate signaling of two soluble ligands or receptors, (c) activate signaling by a ligand mimetic, and (d) delivery across the blood–brain barrier.

Figure 8.2 The bispecific antibody light and heavy chain pairing problems. Coexpression of two light and two heavy chains can result in nine unwanted species besides the desired bispecific antibody. An efficient bispecific platform requires correct light chain pairing with its cognate heavy chain, as well as heterodimeric heavy chain pairing.

Figure 8.3 Examples of bispecific antibodies classified by their mode of production. Antibody domains are colored in red and blue, alternative scaffold domains in green and orange, and the T‐cell receptor domains in brown.

Chapter 09

Figure 9.1 Repositories have been established to manage a diversity of sample types. Compound management facilities may contain millions of different structures but the management of the compounds is similar for each chemical type. Biologics repositories encompass a more diverse range of sample types that may require diverse and unique sample management practices [2, 3].

Figure 9.2 Production of large molecules though complex follows several basic steps from DNA synthesis through purification [8, 9].

Figure 9.3 (a) Core infrastructure for sample management. Registration, inventory management, and request functionality are dispersed into modular systems that can be directly integrated

via

application programming interface (API) or manual processes. (b) In developing systems functions may overlap such that registration of different reagent types are in different registration systems; requests may be integrated into inventory management software.

Figure 9.4 The hierarchy of registration for biologics is similar to that established in compound management: concept, batch/lot/sample.

Figure 9.5 Ancestry or relatedness for biologics is more complex than observed for compounds. Here four examples are shown of how the ancestry of a protein may be tracked from DNA construct to protein.

Figure 9.6 Implementation of automated stores for the management of large molecules.

Figure 9.7 Stability of reagent antibodies stored in bulk at 4°C immediately after purification of the antibody preps contained <5% aggregation. Upon storage aggregation rates trended upward over time but were antibody dependent.

Chapter 10

Figure 10.1 Two‐dimensional gel electrophoresis profile of

E

.

coli

proteins at the end of anti‐CD‐18 blank run fermentation. .

Figure 10.2 Nonlinear dilution of in‐process pool samples for an MAb product in clinical development [6].

Figure 10.3 “Sandwich” capture ELISA assay.

Figure 10.4 Two‐dimensional gel electrophoresis‐SDS‐PAGE gel (a) with Sypro ruby stain and (b) Western blot detection using and anti‐CHO protein (CHOP) antibody.

Figure 10.5 Assessment of HCP clearance during the purification of a mAb. HCPs are detected and quantitated for each process pool and the final drug substance.

Figure 10.6 The detection of protein standards spiked in to mAb were definitiely achieved at 10 ng/mg and some at around 1 ng/mg.

Chapter 11

Figure 11.1 Three main regions can be distinguished in the IgG antibody structure: the Fc region, involved in the dimerization between the two heavy chains (black), and two Fab regions, which result from the interaction between one light chain (gray) and one heavy chain coupled by disulfide bridges. The dots denote glycan moieties. The two Fabs are connected with the Fc

via

the hinge region, in which interchain disulfide bridges between the two heavy chains occur. Glycosylation on a conserved site in the C

H

2 domain is the dominant PTM on IgGs and is required for full effector function. The variable domains (VL, variable light; VH, variable heavy) determine antibody specificity and contain the antigen binding site (CDRs, complementarity determining regions). .

Figure 11.2 A typical molecular assessment workflow encompassing many of the techniques outlined in this chapter are displayed. These techniques can be interchanged with other assays, depending on the biomolecule being tested, the properties the company performing the tests is most interested in, and based upon previous testing used on other bioversions of the molecule of interest. The authors thank Christoph Spiess, Karthik Rajagopal, and Paul McDonald for their help in preparing this figure.

Figure 11.3 An example of a Fab molecule with solubility issues observed during early development. The molecule solubility is pH dependent where it transitions from clear solution (pH 5) to turbid solution (pH 6) to a gel (pH 7) (top panel). The molecule’s solubility also showed salt dependence exhibiting higher solubility in solutions with high NaCl concentration (bottom pane). .

Figure 11.4

In silico

digestions of the antibody or biotherapeutic of interest are performed. Samples are reduced and alkylated and then digested with a panel of proteolytic enzymes. Samples are then cleaned up using a reverse‐phase resin (such as a stage‐tip) and analyzed using LC‐MS/MS. Extracted ion chromatograms corresponding to the masses of predicted peptides, with or without oxidation or deamidation (or both), are selected and quantified by measuring the area under the curve. A summary of these observations is produced allowing researchers to identify “hotspots” of modified residues after a battery of molecular assessment testing has been performed.

Chapter 12

Figure 12.1 Examples of α‐ and β‐anomeric linkages. Provided are the Galα1‐3Gal and Galβ1‐4GlcNAc glycan sequences. GlcNAc,

N

‐acetylglucosamine; Gal, galactose.

Figure 12.2 Examples of high‐mannose, hybrid, and complex N‐linked glycan structures. Man, mannose; GlcNAc,

N

‐acetylglucosamine; NeuAc,

N

‐acetylneuraminic acid; NeuGc,

N

‐glycolylneuraminic acid; Gal, galactose; Fuc, fucose.

Figure 12.3 Mucin‐type O‐linked glycan core structures. Ser, serine; Thr, threonine; Gal, galactose; GalNAc,

N

‐acetylgalactosamine; GlcNAc,

N

‐acetylglucosamine.

Figure 12.4 Reductive amination. Shown is the labeling of

N

‐acetylglucosamine with 2‐aminobenzoic acid (2‐AA) by reductive amination.

Figure 12.5 Derivatization of

N

‐acetylneuraminic acid (NeuAc) with

o

‐phenylenediamine (OPD).

Figure 12.6 Derivatization of

N

‐acetylneuraminic acid (NeuAc) with 1,2‐diamino‐4,5‐methylenedioxybenzene (DMB). BME, β‐mercaptoethanol; HOAc, acetic acid; Na

2

S

2

O

4

, sodium hydrosulfite.

Figure 12.7 Common nomenclature for fragment ions in the mass spectra of peptides. In CID, peptide fragmentation occurs at the amide bond leading to the formation of b‐ and y‐type ions. In ETD, peptide fragmentation occurs at the N–Cα bond resulting in the formation of c‐ and z‐type ions.

Figure 12.8 β‐Elimination/Michael addition reactions. O‐linked glycans are removed from Ser or Thr residues by base‐catalyzed β‐elimination to form dehydroalanine (Dha) or dehydrobutyric acid (Dhb), respectively. Nucleophiles then react with the dehydroamino acids to form stable derivatives. Shown are Michael additions using NH

4

OH and 2‐aminoethanethiol (2‐AET). In the chemical structure of the peptide, R = H for Ser and R = CH

3

for Thr.

Figure 12.9 Edman degradation reaction. Phenyl isothiocyanate (PITC) reacts with the amine group of the N‐terminal amino acid of a peptide. Treatment with a strong acid, like TFA, under anhydrous conditions releases the N‐terminal amino acid as an anilinothiazolinone (ATZ) amino acid, which is then converted to a phenylthiohydantoin (PTH) amino acid by treatment with an aqueous acid solution.

Figure 12.10 ESI‐MS spectrum of an intact antibody. Both antibody Fc glycoforms associated with a particular mass are provided. For example, the antibody glycovariant associated with a mass of 146 225 Da contains a G0F glycan at both Fc glycosylation sites (G0F/G0F). *Masses associated with isobaric glycoforms. All (G1F/G1F) glycoforms are isobaric with (G0F/G2F) glycoforms. In all glycoforms containing “+GlcNAc,” “−GlcNAc,” or “S” in their nomenclature, isomeric forms are possible in which the GlcNAc or NeuAc residue in question can be associated with either of the two glycans that comprise the antibody Fc glycosylation. The structures of the glycans assigned to each glycoform are provided in Table 12.1. Glycan structural assignments are inferred based on a correlation between the observed mass of the glycan moiety and the mass of glycans known to be expressed on CHO‐derived glycoproteins.

Figure 12.11 ESI‐MS spectrum of the heavy chain from the analysis of reduced antibody. The structures of the glycans assigned in the spectrum are provided in Table 12.1. Glycan structural assignments are inferred based on a correlation between the observed mass of the glycan moiety and the mass of glycans known to be expressed on CHO‐derived glycoproteins.

Figure 12.12 Positive ion mode MALDI‐TOF MS analysis of glycans released from a recombinant antibody. The structures of the glycans assigned in the spectrum are provided in Table 12.1. Glycan structural assignments are inferred based on a correlation between the observed mass of the glycan moiety and the mass of glycans known to be expressed on CHO‐derived glycoproteins.

Figure 12.13 CE analysis of APTS‐labeled glycans released from a recombinant antibody. The structures of the glycans assigned in the electropherogram are provided in Table 12.1. Glycan structural assignments are inferred based on a combination of experimental data, including coelution with glycan standards, enzymatic studies, and a correlation between the observed mass of the glycan moiety and the mass of glycans known to be expressed on CHO‐derived glycoproteins or by coelution with glycan standards in CE‐MS experiments. Refer to Section 12.8 for a general description of these methodologies. Glycan assignments were provided by Lynn A. Gennaro. *Resolved glycan structural isomers.

Figure 12.14 HILIC analysis of 2‐AB‐labeled glycans released from a recombinant antibody (expanded view). The structures of the glycans assigned in the chromatogram are provided in Table 12.1. Glycan structural assignments are inferred based on a combination of experimental data, including coelution with glycan standards, enzymatic studies, and a correlation between the observed mass of the glycan moiety and the mass of glycans known to be expressed on CHO‐derived glycoproteins or by coelution with glycan standards in CE‐MS experiments. Refer to Section 12.8 for a general description of these methodologies. Glycan profile and structural assignment provided by Tomasz K. Baginski. *Resolved glycan structural isomers.

Figure 12.15 Preparation and GC–MS analysis of glycan partially methylated alditol acetates (PMAAs). Glycan PMAAs are produced by sequential permethylation, acid hydrolysis, reduction with sodium borodeuteride (NaBD

4

), and acetylation. Provided are the representations of the derivatives produced in the preparation of PMAAs from the Galβ1‐4GlcNAcβ1‐2Manα1 sequence commonly observed in mammalian glycans. Also shown are the masses of some of the characteristic fragment ions observed in the GC–MS analysis of the PMAAs (bottom panel). Ions produced by secondary fragmentation are often observed which result from the loss of formaldehyde (

m

/

z

 = 30), methanol (

m

/

z

 = 32), or acetic acid (

m

/

z

 = 60) from the primary fragment ions.

Figure 12.16 Glycan fragmentation nomenclature. Glycan fragments are labeled using the nomenclature proposed by Domon and Costello [126]. A, B, and C ions are fragment ions that retain the nonreducing terminus of the glycan, whereas X, Y, and Z ions are fragment ions that retain the reducing terminus of the glycan. A and X ions are ions produced by cross‐ring cleavage of two bonds within a constituent monosaccharide. Subscript numbers denote the glycosidic bond cleaved or the monosaccharide involved in cross‐ring cleavage, with reducing terminal fragments (A, B, and C ions) starting with the number 0 and nonreducing terminal fragments (X, Y, and Z ions) starting with the number 1. Superscript numbers on A and X fragment nomenclature denote the bonds within the monosaccharide ring that are cleaved to generate cross‐ring cleavage, with the bonds labeled 0 through 5 starting with the bond between the oxygen and the anomeric carbon.

Figure 12.17

1

H NMR spectrum of a fucosylated fully sialylated biantennary glycan. Provided are the full spectrum (top) and expanded views of the anomeric/H‐2 region (bottom left) and upfield region (bottom right) of the spectrum. Anomeric and Man H‐2 proton signals are labeled based upon the numbering system provided in the glycan structure provided. The Fuc H‐5 and Gal H‐3 proton signals are also observed in the Man H‐2 region and overlap with the signal from the α1,6‐linked Man H‐2 signal. * Methyl protons from isopropanol in the glycan sample. Spectral assignments are based upon published assignments from the

1

H NMR analysis of fully sialylated biantennary glycans [16, 145, 146]. Spectral acquisition and consultation provided by Ken Skidmore.

Guide

Cover

Table of Contents

Begin Reading

Pages

iii

iv

v

xv

xvi

xvii

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

155

156

157

158

159

160

161

162

163

164

165

166

167

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

Analytical Characterization of Biotherapeutics

This edition first published 2017© 2017 John Wiley & Sons, Inc.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.

The right of Jennie R Lill, Wendy Sandoval to be identified as the authors of the editorial material in this work has been asserted in accordance with law.

Registered OfficeJohn Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA

Editorial Office111 River Street, Hoboken, NJ 07030, USA

For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.

Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats.

Limit of Liability/Disclaimer of WarrantyIn view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

Library of Congress Cataloging‐in‐Publication Data

Names: Lill, Jennie R., editor. | Sandoval, Wendy, editor.Title: Analytical characterization of biotherapeutics / edited by Jennie R. Lill, Wendy Sandoval.Description: Hoboken, NJ : Wiley, 2017. | Includes bibliographical references and index. |Identifiers: LCCN 2017013986 (print) | LCCN 2017022128 (ebook) | ISBN 9781119384427 (pdf) | ISBN 9781119384403 (epub) | ISBN 9781119053101 (hardback)Subjects: LCSH: Proteins–Therapeutic use. | Proteins–Analysis. | BISAC: SCIENCE / Chemistry / Analytic.Classification: LCC RM666.P87 (ebook) | LCC RM666.P87 A53 2017 (print) | DDC 615.7–dc23LC record available at https://lccn.loc.gov/2017013986

Cover image: (Background) © Zffoto/Gettyimages;(Illustration) Courtesy of Allison BruceCover design by Wiley

“To Joe and Charlie, thank you for your wonderful smiles and boundless energy”

Jennie R. Lill

“To my daughters, Nicolina and Olivia, who are my daily inspiration”

Wendy Sandoval

List of Contributors

Corey E. BakalarskiDepartments of Microchemistry, Proteomics & Lipidomics and Bioinformatics & Computational BiologyGenentech Inc.South San Francisco, CA, USA

Anne BaldwinDepartments of Biomolecular Engineering and Proteomics & Biological ResourcesGenentech Inc.South San Francisco, CA, USA

Karen BilleciDepartment of Proteomics & Biological ResourcesGenentech Inc.South San Francisco, CA, USA

John B. BriggsDepartment of Protein Analytical ChemistryGenentech Inc.South San Francisco, CA, USA

Natalie CastellanaDigital Proteomics LLCLa JollaandMapp Biopharmaceutical, Inc.San Diego, CA, USA

Paola Di LelloDepartment of Structural BiologyGenentech Inc.South San Francisco, CA, USA

Diego EllermanDepartment of Protein Chemistry and Structural BiologyGenentech Inc.South San Francisco, CA, USA

Adrian GuthalsDigital Proteomics LLCLa JollaandMapp Biopharmaceutical, Inc.San Diego, CA, USA

Robert F. KelleyDrug Delivery DepartmentGenentech Inc.South San Francisco, CA, USA

Denise KrawitzDepartment of Analytical OperationsGenentech Inc.South San Francisco, CA, USA

Jennie R. LillDepartment of Microchemistry, Proteomics & LipidomicsGenentech Inc.South San Francisco, CA, USA

Yichin LiuDepartment of Biochemical and Cellular PharmacologyGenentech Inc.South San Francisco, CA, USA

T. Noelle LombanaDepartment of Antibody EngineeringGenentech Research and Early DevelopmentSouth San Francisco, CA, USA

Patrick LupardusDepartment of Structural BiologyGenentech Inc.South San Francisco, CA, USA

Till MaurerDepartment of Protein Chemistry and Structural BiologyGenentech Inc.South San Francisco, CA, USA

Christopher M. OverallCenter for Blood ResearchUniversity of British ColumbiaVancouver, British Columbia, Canada

Wilson PhungDepartment of Microchemistry, Proteomics & LipidomicsGenentech Inc.South San Francisco, CA, USA

Jason C. RouseDepartment of Analytical Research and DevelopmentPfizerCambridge, MA, USA

Wendy SandovalDepartment of Microchemistry, Proteomics & LipidomicsGenentech Inc.South San Francisco, CA, USA

Justin M. ScheerAntibody Engineering DepartmentBoerhinger Ingelheim, Ridgefield, Connecticut; Department of Protein Chemistry and Structural BiologyGenentech Inc.South San Francisco, CA, USA

Kurt SchroederDepartments of Biomolecular Engineering and Proteomics & Biological ResourcesGenentech Inc.South San Francisco, CA, USA

Lovejit SinghDepartments of Biomolecular Engineering and Proteomics & Biological ResourcesGenentech Inc.South San Francisco, CA, USA

Nestor SolisCenter for Blood ResearchUniversity of British ColumbiaVancouver, British Columbia, Canada

Justin B. SperryDepartment of Analytical Research and DevelopmentPfizerCambridge, MA, USA

Christoph SpiessDepartment of Antibody EngineeringGenentech Research and Early DevelopmentSouth San Francisco, CA, USA

Martin VanderlaanDepartment of Analytical OperationsGenentech Inc.South San Francisco, CA, USA

1Introduction to Biotherapeutics

Jennie R. Lill

Department of Microchemistry, Proteomics & Lipidomics, Genentech Inc., South San Francisco, CA, USA

Abbreviations

ADAs

antidrug antibodies

ADC

antibody–drug conjugate

ADCC

antibody‐dependent cell‐mediated cytotoxicity

CDR

complementary‐determining region

Fab

antigen binding fragment

Fc

cystallizable fragment

NMR

nuclear magnetic resonance

PEG

polyethyleneglycol

PTM

posttranslational modification

1.1 Introduction

Biotherapeutics, also known as biologics, include protein‐based and nucleic acid‐based drugs that are commonly derived by recombinant expression in living organisms although a few are made by chemical synthesis. This book focuses on the characterization of protein‐based biotherapeutics, exploring the various analytical technologies that have enabled in‐depth molecular characterization while discussing current triumphs and limitations.

The first human protein therapeutic derived from recombinant DNA technology was human insulin (Humulin®) created at Genentech, developed by Eli Lilly, and approved by the US Food and Drug Administration (FDA) in 1982. Since that time, major advancements in both recombinant DNA technology and recombinant protein production have contributed to the development of several hundred biotherapeutics [1] including relatively simple molecules such as interferons, insulin, and the human growth hormone to more complexly engineered moieties including ADCs such as trastuzumab emtansine [2] and brentuximab vedotin [3].

Unlike conventional small molecule (chemical) drugs such as aspirin, antibiotics, and various chemo‐therapeutics, the manufacturing process for biotherapeutics is typically far more cumbersome as they are larger compounds with more complex structures and their production can be extremely sensitive to changes in fermentation and environmental conditions. In addition, biotherapeutics are often less stable than many small molecules and can be prone to aggregation [4] or deamidation, oxidation, and other modifications [5]. Since the manufacturing of biotherapeutics is often dependent upon the host cells of living organisms, complex process development is required to ensure reproducible fermentations, isolation, and characterization [6].

1.2 Types of Biotherapeutics and Manufacturing Systems

There are several different types of marketed biotherapeutics including antibody‐based drugs, anticoagulants, blood factors, bone morphogenetic proteins, engineered protein scaffolds, enzymes, Fc (cystallizable fragment) fusion proteins, growth factors, hormones, interferons, interleukins, and thrombolytics (Figure 1.1).

Figure 1.1 Various categories of the main types of biotherapeutics currently marketed.

Source: Carter [7]. Reproduced with permission of Elsevier.

Antibody‐based drugs represent the largest and most rapidly expanding class of biotherapeutics [1]. Figure 1.2 shows the diverse mechanisms by which the antibody structure can be modified to increase its biotherapeutic potential.

Figure 1.2 Monoclonal antibody (mAb) structure can be modified on the basis of the desired mechanism of action. Immunoglobulin G1 (IgG1) is the most effective naturally occurring human IgG isotype at mediating antibody‐dependent cell‐mediated cytotoxicity (ADCC). Glycomodified afucosylated mAbs (part a) (such as Obinutuzumab) demonstrate enhanced binding to IgG Fc receptors (FcγRs) and enhanced ADCC. In addition antibody‐dependent cellular phagocytosis, a process mediated by macrophages, can also occur [8]. Afucosylated mAbs are produced using cell lines that lack the enzymes responsible for fucosylation. Modifying the amino acid sequence of mAb Fc (part b), as was done to produce ocaratuzumab [9], can also result in enhanced binding to FcγRs and enhanced ADCC. For mechanisms of action in which ADCC is not desirable, IgG4 may be a more appropriate isotype, as IgG4 mAbs do not mediate ADCC to the same degree as IgG1 (part c) although this isotype can still engage macrophage effector function via nanomolar affinity binding to FcγRI. Nivolumab, an IgG4 mAb that blocks programmed cell death protein 1 (PD1) on T cells, is one such example. Producing radioimmunoconjugates involves linking the radioisotope to the mAb. A stable linker is most desirable (part d) to limit the leakage of the free radioactive isotope. Conversely, optimal antibody–drug conjugates (ADCs) use a cleavable linker (part e). To avoid nonspecific toxicity, it is desirable for drugs used in ADCs to be cytotoxic once inside the target cell but nontoxic when bound to the mAb in the circulation. Linkers that are pH‐sensitive or enzymatically cleaved are now a standard component of ADCs. Chimeric antigen receptor (CAR) T cells get their specificity from mAb variable regions but are a form of gene, not protein, therapy. They are produced by inserting DNA coding for the mAb variable region fused to DNA coding for signaling peptides into T cells (part f). Some bispecific antibodies lack a functional constant region so that they do not nonspecifically crosslink activating receptors and activate T cells (part g). The lack of a constant region on such constructs results in a short half‐life, thus requiring continuous infusion to achieve the desired exposure.

Source: Weiner [10]. Reproduced with permission of Nature Publishing Group.

Humanized and other chimeric versions of these antibodies now dominate the market [11] and in the past 5 years have accounted for nearly 30% of all approvals. Various antibody isotypes are now being explored to provide a wealth of functional diversity that is present through the various IgG subclasses that can be exploited to improve clinical safety and performance by increasing stability, reducing adverse events, modulating effector functions, and by the engagement of two antigens by a single antibody [8]. Several variants that have been Fc engineered for reduced effector function have entered the clinic, for example, Eculizumab, a novel engineered IgG isotype, IgG2m4, with reduced Fc functionality. IgG2m4 is engineered based on the IgG2 isotype with four key amino acid residue changes derived from IgG4 (H268Q, V309L, A330S, and P331S). This antibody was demonstrated to have an overall reduction in complement and Fc gamma receptor binding in in vitro binding analyses while maintaining the normal in vivo serum half‐life in rhesus [12].

Biosimilars (biologically identical antibodies, for example) and so‐called biobetters (moieties with improved properties such as pharmacodynamic (PD) and pharmacokinetic (PK) readouts, higher potency, longer half‐lives, and less immunogenicity, for example) are also starting to emerge, which presents new challenges in terms of testing for the presence of liabilities such as degradative properties, changes in immunogenicity through addition of novel contaminant proteins from new manufacturing processes, and so on. New formats such as glucagon‐like peptide 1GLP fused proteins, for example, Eperzan (albiglutide) [13], and PEGylated proteins such as Plegridy (e.g., peginterferon beta‐1a) [14] offer improved PK or PD properties but also increased analytical challenges due to their larger masses and increased heterogeneity.

Typically, expression of non‐mAb biotherapeutics has been performed in Escherichia coli or a noneukaryotic system. This has many advantages for biotherapeutics that are not reliant on PTMs for their optimal activity. Over the years, however, there has been a gradual increase in the prevalence of mammalian expression systems. Of the mammalian expression systems, the Chinese hamster ovary (CHO) cell‐based model (reviewed by Krawitz and Sandoval in Ref. [11]) remains the most employed expression system with a smaller percentage of therapeutics manufactured in other mammalian cell lines such as the murine myeloma line, NSO, and baby hamster kidney cells [15, 16]. Nonmammalian eukaryotic expression systems such as yeast [17] are also utilized, each again presenting their own challenges with regard to the correct PTM of the protein, occasionally adding to adverse properties [18, 19].

More recently transgenic animal production systems (e.g., expression of recombinant products in the animals’ milk [19, 20], rabbits, and goats) have been explored as a means of biopharmaceutical production although to date there are many challenges associated with this type of biotherapeutic production with few benefits. Throughout this book the challenges of characterizing both the biotherapeutic moiety itself and the contaminant proteins such as CHO‐derived proteins are discussed.

1.3 Types of Analyses Performed

Throughout this book a variety of analytical procedures are described. Many of them have been implemented for characterizing biotherapeutic molecules for as long as these moieties have existed. Others have evolved as the need arises. One such example of developing such sets of tools to answer a newly arisen problem is for the de novo sequencing of antibodies [21, 22]. Occasionally antibodies are discovered that are of great interest for preclinical testing, for which the cDNA or any genetic information is not available. In these scenarios, researchers have to sequence antibodies at the protein level, one amino acid at a time, and then reverse engineer the antibodies to the nucleotide level. In Chapter 6 Castellana and Guthals provide technical details and review the innovative approaches employed to quickly gain sequence information through a de novo approach.

As well as sequence information at the amino acid level, PTM profiling is also an important element in characterizing biotherapeutics [23]. There are a plethora of cotranslational modifications and PTMs that play key roles in the folding of proteins, in their secretion, and in their ultimate stability and effector functionality in vivo.

Glycosylation is important both for antibody secretion by B‐cells and for in vivo antibody effector function. Glyco‐engineering is a rapidly growing field, whereby glycosylation sites and composites are engineered to produce antibodies with specific glycoforms which may have an effect on therapeutic efficacy. Obinutuzumab (Gazyva®) [24], for example, is a humanized therapeutic monoclonal antibody that binds to an epitope on the B cell antigen, CD20. This antibody is engineered in a platform that allows control of the proteins’ glycosylation, in this case the platform enforces the overexpression of two glycosylation enzymes MGAT3 and the golgi mannosidase 2. This results in the generation of antibodies with bisected nonfucosylated sugars, thereby increasing the antibodies’ ability to activate natural killer cells. This means that Obinutuzumab can induce cell death through a dual mechanism of action, both by the antibody directly binding to B cells and by antibody‐mediated cytotoxicity by recruiting the immune system to attack B cells. Some types of glycosylation are sometimes not beneficial. For example, Cetuximab, a chimeric mouse–human IgG1 monoclonal antibody against the epidermal growth factor receptor (EGFR) approved for use in colorectal cancer and squamous‐cell carcinoma of the head and neck has a high prevalence of hypersensitivity reactions which has been attributed to cross reactivity to a V domain glycosylation site. In some patients severe adverse events have been observed including anaphylaxis, which was found to be due to the generation of patient‐specific antibodies to the galactose‐alpha‐1,3‐galactose modification [25].

Protein terminal modifications have the effect of modifying a protein’s function, half‐life, or cellular localization. Pyroglutamate formation, for example, is a highly prevalent modification whereby glutamine and glutamate at the N termini of recombinant monoclonal antibodies can cyclize spontaneously to pyroglutamate (pE) in vitro [26]. Proteolytic processing is also an irreversible modification that affects the vast majority of proteins, often with great functional consequences. Intracellular proteolytic processing has distinct effects on the functionality of proteins and can either abrogate or antagonize function, modify half‐life, or also determine cellular localization. During protein synthesis, manufacturing, purification, and storage proteolysis events can occur, thereby changing a protein’s functionality or stability [27]. Either through direct mass spectrometric analysis as reviewed in Chapter 2, through a variety of historic analytical techniques such as gel electrophoresis or Edman degradation, or by employing a variety of new biochemical‐based methodologies for determining the termini of recombinant proteins as reviewed in Chapter 3, the determination of proteolytic processing remains a key analytical need for the characterization of biotherapeutic moieties.

Beyond linear sequence determination, structural analyses are also instrumental in the overall characterization of biotherapeutics. The biomolecular architecture is a vital component in dictating the specificity and overall efficacy of therapeutic proteins. The higher order structure (HOS) of a protein includes the secondary, tertiary, and quaternary structures of a protein that are required for its function. There is a diverse range of biophysical methods including circular dichroism, isothermal calorimetry, which are available for the characterization of a protein HOS, each of them with associated benefits and limitations. Related to conformational analysis is structural analysis as it pertains to epitope and paratope mapping. Again, several well‐established techniques such as nuclear magnetic resonance (NMR) [28] and X‐ray crystallography as well as some newer techniques such as mass spectrometric‐based structural tools [29] including hydrogen deuterium exchange are described in Chapters 4 and 5.

1.4 Future perspectives

Nature has provided us with various types of protein scaffolds to explore as frameworks for building new types of biotherapeutics and there is a growing field of using these scaffolds as alternatives to antibodies [30]. Each of these types of engineered molecular structures offers new advantages in terms of stability and specificity. One example of this is the cystine knot mini proteins/peptides (knottins); these are peptide‐based alternative molecules to monoclonal antibodies which are raised/designed against tumor‐associated receptors and other antigens of interest. Knottins contains a disulfide‐bonded core that exhibits a high level of resistance to proteolysis and increased thermal stability. Knottins emerged as an attractive molecular candidate for drug development as they fill the niche between small molecule drug design and protein biologics. Knottins have the potential to bind clinical targets with both high selectivity and affinity [31]. There are several naturally occurring knottins that have been approved as biotherapeutics for the treatment of pain [32] and irritable bowel syndrome and for tumor imaging purposes [33].

Elucidating disulfide bonding patterns of any biomolecules, but in particular a structure which relies on disulfide bonding patterns for their folding, stability, and activity, is an important part of molecular characterization. A variety of techniques can be employed for doing this from simple intact molecular weight measurement to more complex top‐down proteomic protocols [34] and these methodologies continue to mature as more of these types of molecules emerge on the market.

Another growth area for biotherapeutics is increasing the molecules’ in vivo half‐life. For therapeutics that involves frequent or uncomfortable delivery, for example, injectable ocular therapeutics, or to make drugs that have poor PD properties more tolerable, there is a strong drive to create molecules that have increased in vivo