mAbs
`
`ISSN: 1942-0862 (Print) 1942-0870 (Online) Journal homepage: https://www.tandfonline.com/loi/kmab20
`
`Developability studies before initiation of process
`development
`
`Xiaoyu Yang, Wei Xu, Svetlana Dukleska, Sabrina Benchaar, Selina Mengisen,
`Valentyn Antochshuk, Jason Cheung, Leslie Mann, Zulfia Babadjanova, Jason
`Rowand, Rico Gunawan, Alexander McCampbell, Maribel Beaumont, David
`Meininger, Daisy Richardson & Alexandre Ambrogelly
`
`To cite this article: Xiaoyu Yang, Wei Xu, Svetlana Dukleska, Sabrina Benchaar, Selina Mengisen,
`Valentyn Antochshuk, Jason Cheung, Leslie Mann, Zulfia Babadjanova, Jason Rowand, Rico
`Gunawan, Alexander McCampbell, Maribel Beaumont, David Meininger, Daisy Richardson &
`Alexandre Ambrogelly (2013) Developability studies before initiation of process development,
`mAbs, 5:5, 787-794, DOI: 10.4161/mabs.25269
`To link to this article: https://doi.org/10.4161/mabs.25269
`
`Copyright © 2013 Landes Bioscience
`
`View supplementary material
`
`Published online: 07 Jun 2013.
`
`Submit your article to this journal
`
`Article views: 2651
`
`Citing articles: 51 View citing articles
`
`Full Terms & Conditions of access and use can be found at
`https://www.tandfonline.com/action/journalInformation?journalCode=kmab20
`
`Lassen - Exhibit 1012, p. 1
`
`

`

` REPORT
`mAbs 5:5, 787–794; September/October 2013; © 2013 Landes Bioscience
`
`REPORT
`
`Developability studies before initiation
`of process development
`Improving manufacturability of monoclonal antibodies
`
`Xiaoyu Yang,1 Wei Xu,1 Svetlana Dukleska,1 Sabrina Benchaar,2 Selina Mengisen,1 Valentyn Antochshuk,3 Jason Cheung,3
`Leslie Mann,1 Zulfia Babadjanova,1 Jason Rowand,1 Rico Gunawan,1 Alexander McCampbell,4 Maribel Beaumont,2
`David Meininger,2 Daisy Richardson1 and Alexandre Ambrogelly1,*
`
`1BioProcess Development; Merck Research Laboratories; Union, NJ USA; 2Discovery Biologics; Merck Research Laboratories Palo Alto; Palo Alto, CA USA;
`3BioProcess Development; Merck Research Laboratories; Summit, NJ USA; 4Molecular biomarkers; Merck Research Laboratories; West Point, PA USA
`
`Keywords: developability, discovery, stability, analytics, process control, development, forced degradation, CQA, QbD
`
`Monoclonal antibodies constitute a robust class of therapeutic proteins. Their stability, resistance to stress conditions
`and high solubility have allowed the successful development and commercialization of over 40 antibody-based drugs.
`Although mAbs enjoy a relatively high probability of success compared with other therapeutic proteins, examples
`of projects that are suspended due to the instability of the molecule are not uncommon. Developability assessment
`studies have therefore been devised to identify early during process development problems associated with stability,
`solubility that is insufficient to meet expected dosing or sensitivity to stress. This set of experiments includes short-term
`stability studies at 2–8°C, 25°C and 40°C, freeze-thaw studies, limited forced degradation studies and determination of
`the viscosity of high concentration samples. We present here three case studies reflecting three typical outcomes: (1) no
`major or unexpected degradation is found and the study results are used to inform early identification of degradation
`pathways and potential critical quality attributes within the Quality by Design framework defined by US Food and Drug
`Administration guidance documents; (2) identification of specific degradation pathway(s) that do not affect potency of
`the molecule, with subsequent definition of proper process control and formulation strategies; and (3) identification of
`degradation that affects potency, resulting in program termination and reallocation of resources.
`
`Introduction
`
`Development of a therapeutic protein is a long and costly process
`that can take over a decade from discovery to commercialization.
`Although therapeutic biologics generally have a higher probabil-
`ity of success than their small molecule counterparts,1 the rate of
`attrition remains substantial. While decisions to terminate proj-
`ects can be based on the competitive landscape and commercial
`opportunities, projects are also terminated for technical reasons,
`including unsuitable safety profile, lack of efficacy in human,
`instability of the molecule or formulation, poor expression and
`purification issues. To reduce the risk associated with a given
`project, the sooner these technical challenges are identified, the
`sooner appropriate control strategies can be put in place. When
`irreducible challenges are identified, the project can be termi-
`nated and resources diverted to the development of a more prom-
`ising molecule. Derisking process development early enough is
`critical to success because challenges not identified early enough
`may lead to expensive and time-consuming remediation steps
`later in the development of a given program. Determination of
`the safety profile is usually performed during preclinical studies
`
`and Phase 1 clinical studies in human, which is well into the
`development path. Likewise, the true assessment of a molecule
`efficacy is obtained during Phase 2/3 clinical studies.
`In contrast with safety and efficacy assessments, a number
`of technical hurdles can be evaluated early during development
`programs. To identify technical challenges, we devised a compre-
`hensive set of developability experiments and applied the strategy
`to a number of monoclonal antibodies (mAbs) currently in our
`pipeline. This strategy includes a series of in silico and experi-
`mental approaches that we grouped under the Developability
`Assessment term. In silico work includes structure sequence
`analysis based on primary sequence alignment and molecular
`model; prediction of possible degradation pathways is based on
`prior experience and on published literature. This constitutes the
`first step toward establishing potential critical attributes within
`the Quality by Design (QbD) paradigm encouraged by the US
`Food and Drug Administration.2 The experimental part of the
`developability assessment includes short-term stability studies at
`various temperatures, freeze-thaw studies and limited forced deg-
`radation studies. This series of studies determine the biochemical
`(e.g., sensitivity to oxidation and deamidation) and biophysical
`
`*Correspondence to: Alexandre Ambrogelly; Email: alexandre.ambrogelly@merck.com
`Submitted: 04/28/13; Revised: 05/31/13; Accepted: 06/03/13
`http://dx.doi.org/10.4161/mabs.25269
`
`www.landesbioscience.com
`
`mAbs
`
`787
`
`Lassen - Exhibit 1012, p. 2
`
`

`

`Conditions
`
`Initial
`
`Table 1. Developability stability schedule
`t = 6
`t = 3
`t = 2
`t = 0.5
`t = 1
`mo
`mo
`mo
`mo
`mo
`/
`/
`/
`/
`X
`−80°C
`X
`X
`X
`X
`X
`2–8°C
`X
`X
`X
`X
`X
`25°C
`X
`X
`X
`X
`X
`40°C
`/ No samples staged at these conditions/timepoints. X Samples staged
`and analyzed at these conditions/timepoints. Details of the assays used
`during stability are reported in Table S1.
`
`X
`
`(e.g., unfolding and formation of large molecular weight spe-
`cies) stability profiles of the molecule of interest. Determination
`of maximum solubility and associated viscosity are also essential
`for molecules destined to be formulated at high concentrations
`for subcutaneous injections. Developability studies offer the first
`opportunity to assess the technical viability of a project at the
`industrial scale and function as a bridge between discovery and
`process development activities. We present in this report three
`case studies and include data generated during developability
`assessment of antibody molecules.
`
`Results
`
`General strategy for developability studies. Developability stud-
`ies are considered a derisking activity designed to provide under-
`standing of the colloidal properties and prevalent degradation
`pathways of a molecule early in the development process, thereby
`allowing the analytical strategy to be tailored to the molecule
`of interest. To be relevant, developability studies should be per-
`formed on material produced with a stable cell line of the same
`host as the one that will be used later for process development;
`when possible, use of the final stable clone is preferable.
`The most widespread host for expression of mAbs is the
`Chinese hamster ovary (CHO) cell line; however, a developabil-
`ity strategy can be applied to material produced by any expression
`system. The recombinant protein is produced via an upstream
`early cell culture process and purified by a downstream, 2–3 col-
`umns standard process. It is important to note that, at this stage
`of development, the formulation has not been optimized and the
`protein matrix might be suboptimal. Keeping these limitations in
`mind, developability studies are aimed at detecting gross changes
`in the protein that could hinder the stability or potency of the
`molecule.
`Sequence alignment and molecular modeling. Developability
`assessment starts with analysis of the sequence of the protein
`entering the development phase. Some of this work has likely been
`done in the discovery space. For example, potentially problematic
`residues such as methionine, asparagine or aspartic acid residues
`localized in exposed region of the mAb, including the comple-
`mentarity-determining regions (CDRs), may have been removed.
`Relevant knowledge and experience generated during the discov-
`ery phase, such as whether the molecule was generated through
`phage display or hybridoma cells, should be taken into account.
`The first examination of the sequence during developability
`
`assessment is therefore aimed at identifying potential hot spots
`for degradation that remain in the sequence. Sequence alignment
`can be performed by any software available on the market. To
`complete the analysis of the primary sequence, building a molec-
`ular model may help to visualize residues exposed to the solvent,
`which are susceptible to degradation. The software we currently
`use for this exercise is MOE (CCG, Montreal, Canada).
`Short-term research stability studies and freeze/thaw cycles. To
`gain a first impression about how the molecule will degrade,
`short-term stability studies are performed. Material is staged
`at −80°C, 2–8°C, 25°C and 40°C for up to 6 mo (Table 1).
`Samples are removed at different time points and analyzed by size
`exclusion (SEC) HPLC to detect the formation of high molecular
`weight species and by ion exchange (IEX) HPLC to measure the
`effect of stress on distribution of charges at the surface of the
`molecule due to the deamidation and isomerization of asparagine
`and aspartic acid residues. Samples are also analyzed by reducing
`and non-reducing sodium dodecyl sulfate capillary electrophore-
`sis (CE-SDS) to detect proteolytic cleavage of the heavy or light
`chains, peptide mapping to quantify the formation of post-trans-
`lational modifications formed during stress and a potency test,
`which is most commonly in a binding ELISA format in the very
`early stage of the program development.
`Propensity of the recombinant protein of interest to form
`aggregates is gauged during freeze/thaw cycles. The protein is
`subjected to five cycles. Aliquots are analyzed by SEC-HPLC,
`SEC-multi-angle laser light scattering and dynamic light scatter-
`ing at each of the five cycles.
`Limited stress conditions. To complete the initial assessment
`of the stability of the molecule, a series of stress conditions are
`applied (Table 2). These conditions include standard stresses
`suggested by regulatory agencies, such as exposure to low and
`high pHs, light and oxidative reagents.3 More emphasis may be
`put on some of these stress conditions if a potential degradation
`pathway has been identified by molecular modeling or during
`early experience with the material. The stresses are likely to trig-
`ger deamidation, isomerization, aggregation and oxidation of
`amino acid side chains. These degradation pathways are the most
`commonly described for mAbs.4 A direct measurement of the
`effect of the degradation on potency is also taken.
`Determination of viscosity and ability to concentrate formulated
`materials (surrogate for solubility studies). High concentrations of
`protein are usually required when mAb therapeutics are admin-
`istered via subcutaneous injection. To verify whether the drug
`candidate can achieve a target concentration, the protein is con-
`centrated to 100–200 mg/mL in standard buffers. Viscosity of
`the high concentration solutions is measured to evaluate purifica-
`tion and delivery device strategies.
`mAb A case study: No major or unexpected degradation.
`The first case study concerns an IgG1 antibody, termed mAb
`A. Sequence structure analysis of the CDRs did not reveal any
`particularly exposed asparagine or aspartic residues in an amino
`acid context prone to deamidation or isomerization (i.e., located
`directly upstream from a glycine, serine or proline residues).7
`CDRs did not contain exposed methionine or unpaired cys-
`teine residues. The short-term stability of early representative
`
`788
`
`mAbs
`
`Volume 5 Issue 5
`
`Lassen - Exhibit 1012, p. 3
`
`

`

`Table 2. Typical limited forced degradation conditions
`Tests
`Duration
`Condition
`Step
`UV, SEC, IEX, peptide mapping, potency
`48 hr
`Low pH
`Adjust pH with HCl to 3.0
`UV, SEC, IEX, peptide mapping, potency
`48 hr
`High pH
`Adjust pH with NaOH to 9.0
`UV, SEC, IEX, peptide mapping, potency
`N/A
`Light stress
`1X ICH (UV, Visible)
`UV, SEC, IEX, peptide mapping, potency
`2–5 h
`Methionine oxidation
`0.05% tBHP
`Measure A280, A320 and DLS before and after pH adjustment to pH 5.5
`N/A
`pH jump
`Coordinate with purification area
`tBHP, tert-Butyl hydroperoxide a reagent.5 ICH, International Conference on Harmonisation; details on guidance for Light stress can be found in ICH
`guideline Q1B.6 Details of the assays used during forced degradation are reported in Table S1.
`
`Figure 1. Overlay of IEX-HPLC chromatograms of mAb A samples stored for one month (A) and three
`months (B) at −80°C, 2–8°C, 25°C and 40°C. Overlay of IEX-HPLC chromatograms of mAb A samples chal-
`lenged with high and low pHs for 48 h at 40°C (C) and light induced (0.2xICH) stress (D).
`
`material was evaluated for up to
`three months at 2–8°C, 25°C and
`40°C via a battery of assays. IEX-
`HPLC profiles after one and three
`months are shown in Figure 1A and
`1B. Decrease of relative peak area of
`the main and basic species (eluting
`after the main) and increase of the
`relative peak area of the acidic vari-
`ants (eluting before the main) was
`noted for samples stored at 40°C. A
`decrease of basic variants is typical
`for mAbs; temperature accelerates
`the cyclisation of N-terminus gluta-
`mine into pyroglutamic acid, a post-
`translational modification common
`to many therapeutic and endogenous
`human mAbs.8 Likewise, an increase
`of acidic variants upon exposure to
`elevated temperature is a common
`occurrence that may reflect deami-
`dation of asparagine residue, isom-
`erization of aspartic residues into
`isoaspartic or structural changes
`affecting the overall surface charge distribution of the molecule.8
`In the present case, deamidation was only detected at conserved
`position 385 within the Fc portion of the mAb and did not
`exceed 10% (Table 3). Limited forced degradation studies at low
`and high pH (3.0 and 9.0) after 48 h at 40°C, resulted in pH
`dependant degradation routes: exposure to high pH resulted in
`an increase of acidic variant (and deamidation of position 385),
`while low pH resulted in the formation of basic species not iden-
`tified by peptide mapping (Fig. 1C, Table 3). Exposure to other
`forms of stress was performed and did not result in unexpected
`alterations of the primary, secondary or tertiary structure of the
`molecule. Exposure to UV/visible light following ICH guide-
`lines, for instance, resulted in the formation of a basic species
`reflecting oxidation of conserved Fc methionines 253 and 429
`(Fig. 1D, Table 3). Oxidation of the conserved Fc methionine
`residues under these conditions was expected and documented
`in the literature.4 While developability studies of mAb A did not
`identify major chemistry, manufacturing and control (CMC)
`challenges, the outcome of these studies provided information
`about primary degradation pathways and potential critical qual-
`ity attributes (CQA) for the molecule.
`
`5°C
`25°C
`40°C
`
`5°C
`25°C
`40°C
`
`4%
`4%
`4%
`
`5%
`5%
`6%
`
`8%
`8%
`8%
`
`Table 3. Relative amounts of PTMs in mAb A by LC-MS peptide mapping
`Storage Temperature
`Met 253 ox
`Met 429 ox
`N385D
`Initial
`7%
`4%
`4%
`1 mo time point
`8%
`8%
`9%
`3 mo time point
`4%
`8%
`5%
`7%
`4%
`21%
`Limited forced degradation
`14%
`8%
`4%
`pH 3.0/40°C
`7%
`8%
`4%
`pH 9.0/40°C
`5%
`22%
`17%
`Photostress
`6%
`7%
`4%
`Dark Ctrl
`Met253ox and Met429 ox refer to methionine oxidation position 253
`and 429, respectively. N385D refers to relative amount of aspartic acid
`at position 385. Photostress conditions were 0.2x ICH, Dark Ctrl was the
`corresponding control.
`
`www.landesbioscience.com
`
`mAbs
`
`789
`
`Lassen - Exhibit 1012, p. 4
`
`

`

`Figure 2. Overlay of IEX-HPLC chromatograms of mAb B samples stored for one month (A) and three months (B) at −80°C, 2–8°C, 25°C and 40°C. Evo-
`lution during stability studies of relative amount of deamidation at HCAsn55 (N55, blue trace) and mAb B potency (red trace) (C). Overlay of IEX-HPLC
`chromatograms of mAb B samples stored for one month at 40°C at pH 4.5 and pH 6.0 (D). Thermal unfolding profile ofsolution of mAb B (E).
`
`mAb B case study: Identification of degradation pathways
`informs appropriate control strategy. The second example of
`developability studies involves a S228P hinge-modified IgG4
`molecule, termed mAb B. Scanning of the primary sequence
`for exposed asparagine, aspartic acid, tyrosine, tryptophan and
`methionine residues did not predict the presence of particularly
`sensitive hot spots. Results from the short-term stability studies,
`however, show a collapsed IEX-HPLC profile after three months
`at 40°C and a significant increase of acidic variants at 25°C and
`essentially no changes when stored at 2–8°C or frozen (Fig. 2A and
`B). No increase of aggregates or fragments resulting from chemi-
`cal proteolysis was detected by SEC-HPLC (data not shown).
`Peptide mapping and LC-MS analysis indicated that deamida-
`tion of heavy chain Asn55 exceeded 50% and thus was largely
`responsible for the instability of the molecule (Table 4). Other
`expected post-translational modifications, such as oxidation of the
`
`conserved Fc methionine residues were also detected, albeit to low
`levels. Testing of relative potency by binding ELISA showed that,
`in spite of the chemical degradation, mAb B retained its complete
`ability to bind its target antigen (Table 4, Fig. 2C). Lowering the
`pH of the mAb B formulation from 6.0 to 4.8 reduced greatly the
`amount of deamidation after one month at 40°C (Fig. 2D). For
`this particular molecule, deamidation at Asn55 was recognized as
`a potential CQA. A proper control strategy involving monitoring
`deamidation during upstream and down stream process develop-
`ment and manufacturing was put in place. Efforts toward defin-
`ing an adequate formulation and storage conditions of the drug
`substance were made to ensure protein stability and the longest
`possible product shelf-life. mAb B’s desirable commercial drug
`product form is a highly concentrated solution for subcutaneous
`injection. Part of the developability exercise is to verify that the
`protein can remain in solution at concentrations above 100 mg/
`
`790
`
`mAbs
`
`Volume 5 Issue 5
`
`Lassen - Exhibit 1012, p. 5
`
`

`

`−80°C
`5°C
`25°C
`40°C
`
`3.7%
`3.4%
`3.8%
`4.1%
`
`100%
`100%
`103%
`101%
`
`113%
`108%
`94%
`
`M250ox M426ox
`N55D
`1 mo time point
`21.6%
`10.1%
`21.7%
`10.0%
`23.9%
`10.8%
`36.4%
`11.7%
`3 mo time point
`3.1%
`21.1%
`14.7%
`5°C
`2.7%
`28.1%
`16.7%
`25°C
`3.5%
`55.6%
`18.6%
`40°C
`N55D refers to relative amount of aspartic acid at position 55 on the
`heavy chain. Met248ox and Met250ox refer to methionine oxidation
`positions 248 and 250. Potency by binding ELISA is expressed as % of
`reference standard.
`
`Table 4. Potency and Relative amounts of PTMs in mAb B by LC-MS
`peptide mapping
`Storage Temperature
`
`Potency
`
`mL and such a solution has viscosity that would allow develop-
`ment of prefilled syringe image and autoinjector devices. mAb
`B was successfully concentrated to 190 mg/mL and remained in
`solution. At these concentrations, viscosity remained under 19 cP,
`which allows development of purification processes and prefilled
`syringe image (Table 5). Additionally, mAb B has demonstrated
`good thermodynamic stability (onset of the first transition above
`50°C) and thermal profile with three unfolding transitions at
`63, 68 and 80°C, respectively (Fig. 2E, Table 5), which is typi-
`cal for multidomain proteins. Developability studies of mAb B
`highlighted primary degradation pathways and potential criti-
`cal attributes. It also identified CMC issues addressable with an
`appropriate manufacturing and analytical control strategy.
`mAb C case study: Identification of degradation pathways
`leads to project termination and reallocation of resources. The
`third case study concerns an IgG2 molecule, termed mAb C.
`Analysis of the primary sequence of the molecule showed the pres-
`ence of an asparagine-glycine (Asn33-Gly34) sequence in one of
`the CDRs of the antibody light chain. Analysis of the structural
`context of these residues in the molecular model showed that
`Asn33 is exposed to the solvent and located in a flexible loop. The
`Asn-Gly sequence has been documented as a site particularly sen-
`sitive to facile deamidation to aspartic acid and isomerization to
`isoaspartic acid.7 IEX-HPLC chromatograms of mAb C samples
`showed a series of three charge variant peaks eluting at about 8, 12
`and 16 min (Fig. 3A and B). These peaks were first believed to be
`associated with the three different hinge region disulfide isomers
`commonly found in IgG2s with kappa light chain.9 Isolation and
`characterization of the charge variants demonstrated, however,
`that the two acidic variants (eluting at 8 and 12 min) were mAb
`C molecules with one or both of the Asn33 deamidated (data
`not shown). Analysis by IEX-HPLC of the short-term stability
`at one and three months time points showed a rapid increase of
`acidic variants when the molecule was stored at 25°C and 40°C
`(Fig. 3A and B). Under the same storage conditions and at the
`same time points, the molecule showed only minimal increase of
`high molecular weight species by SEC-HPLC (data not shown).
`Consistent with the increase of the acidic variants peaks, analysis
`of the stability samples by peptide mapping showed dramatic for-
`mation of aspartic acid and isoaspartic acid at light chain position
`33 (Table 6). Deamidation of light chain Asn33 reached 92.6%
`at 40°C after three months and correlated with mAb C loss of
`potency (Table 6, Fig. 3C). Other post-translational modifica-
`tions were also detected: oxidation of the conserved methionine
`248 in the Fc domain and limited cyclization of the glutamic acid
`residue located at heavy chain position 1 (Table 6). These later
`modifications are well-documented and shown to have no effect
`on the potency of mAbs.4,8 Asn33 deamidation can be controlled
`by optimizing the formulation; lowering of the pH reduced deam-
`idation at elevated temperature over time (Fig. 3D). However,
`stabilization of the molecule in vitro is not sufficient. Ex vivo and
`in vivo rhesus monkey plasma stability studies indicated that deg-
`radation of mAb C drug candidate occurs in plasma (Fig. 3E
`and F). Titration of mAb C by anti-kappa light chain and anti-
`gen ligand-capture ELISAs, in serum samples drawn from rhe-
`sus monkeys, showed a time-dependent separation of the PK and
`
`Table 5. Viscosity and thermal melt of mAb B
`Thermal transitions (°C)
`Onset
`Tm1
`Tm2
`Tm3
`
`Concentration
`
`Viscosity (cP)
`
`19
`190
`2
`60
`Tm1, Tm2, Tm3 refer to melting temperature transition 1, 2, 3 shown in
`thermal melt curves (Fig. 2E).
`
`68
`
`80
`
`56
`
`63
`
`antigen binding curves, suggesting a loss of ligand binding capac-
`ity consistent with our in vitro observations (Fig. 3F). To remain
`viable as a product candidate, mAb C had to be reengineered by
`replacing Asn33 by an alternative amino acid without altering
`affinity for the target antigen. Developability studies of mAb C
`led to the identification of major CMC challenges that could be
`addressed only via re-engineering of the molecule.
`
`Discussion
`
`Developability studies are an important link between drug dis-
`covery and process development activities. After expression of a
`proposed therapeutic biologic in the expected host organism, early
`assessment of the molecule’s stability is the first occasion to gather
`knowledge on it to determine potential CQAs as encouraged by
`the QbD paradigm. We presented three cases of developability
`studies, one for each of the typical therapeutic mAb frameworks,
`that correspond essentially to three scenarios encountered during
`developability studies. For the first example, short-term stabili-
`ties, freeze-thaw and limited forced degradation studies did not
`uncover unexpected sources of concerns about the stability of the
`molecule. In this particular case, developability studies showed
`that the molecule’s degradation pathways conformed to those
`of standard IgGs. The second case study provides an example
`of a therapeutic mAb undergoing degradation without affecting
`potency. In this case, a proper analytical strategy is put in place
`to monitor the degradation detected. Proper process controls are
`also instituted to limit the degradation during manufacturing.
`
`www.landesbioscience.com
`
`mAbs
`
`791
`
`Lassen - Exhibit 1012, p. 6
`
`

`

`Figure 3. Overlay of IEX-HPLC chromatograms of mAb C samples stored for one month (A) and three months (B) at −80°C, 2–8°C, 25°C and 40°C.
`Evolution during stability studies of relative amount of deamidation at HCAsn33 (N33, blue trace) and mAb B potency (red trace) (C). The bar graph
`represents the relative area of acidic variants after one month at 40°C in six different formulations (D). Deamidation at Asn33 in rhesus monkey serum
`stability studies (E): TIC of the doubly charged ion of peptide 25–39 at day 0, 7, 14 and 21: The peak at ~33–34 min is the early eluting non deamidated
`parent peptide (Asn33), the two peaks eluting at 34–36 and 35–37 min are the deamidated product peptides (iso-Asp33 and Asp33, respectively). At
`all-time points the ratio of deamidated product peptides IsoAsp/Asp is the same, ~1:5 throughout the whole experiment. Titration of mAb C in rhesus
`monkeys serum by an anti-kappa light chain (PK curve) and antigen ligand-capture (Antigen curve) ELISAs after second injection (F).
`
`Lastly, formulation studies are put in place to ensure that the deg-
`radation does not occur during storage and does not compromise
`the shelf-life of the drug product. The last case study corresponds
`to the case where stability studies detect a major problem with
`the molecule’s stability and efficacy, leading to termination of the
`project and reallocation of resources.
`
`Materials and Methods
`
`Materials. mAb A, mAb B and mAb C are humanized IgG1,
`IgG4 and IgG2, respectively. All molecules have λ light chains.
`
`These Merck proprietary mAbs are expressed in CHO cells and
`purified using three chromatographic steps (protein A, anion
`exchange and cation exchange). Formulations included histi-
`dine, acetate and typical excipients such as sucrose and PS80;
`pHs ranged from 5.0 to 6.0.
`Ion exchange HPLC (IEX-HPLC). Ion exchange HPLC for
`mAbs A, B and C is run on the Dionex WCX-10 column and
`uses either a citrate or Mes-based mobile phase A and a phos-
`phate buffer with or without added NaCl as mobile phase B.
`Gradient is adjusted to suit the needs of each molecule.
`
`792
`
`mAbs
`
`Volume 5 Issue 5
`
`Lassen - Exhibit 1012, p. 7
`
`

`

`5°C
`25°C
`40°C
`
`3.4%
`5.2%
`12.4%
`
`30.3%
`44.1%
`64.5%
`
`Table 6. Potency and Relative amounts of PTMs in mAb C by LC-MS peptide mapping
`Storage Temperature
`N33D
`N33isoD
`N33D+N33isoD
`Initial
`3.5%
`29.2%
`32.6%
`1 mo time point
`33.6%
`49.2%
`76.9%
`3 mo time point
`96%
`1.5%
`3.2%
`30.1%
`26.6%
`3.5%
`5°C
`28%
`2.2%
`5.8%
`62.6%
`56.2%
`6.4%
`25°C
`5%
`6.1%
`21.7%
`92.6%
`68.5%
`24.1%
`40°C
`N33D and N33IsoD refer to relative amount of aspartic acid and isoaspartic acid at position 33, respectively. Total deamidation at position 33 is indi-
`cated as the sum of (N33D and N33IsoD). Met248ox refers to methionine oxidation position 248. HC pE1 refers to the proportion of pyroglutamic acid
`at HC position1. Potency by binding ELISA is expressed as % of control (Initial, indicated as N/A).
`
`Met248 ox
`3.2%
`
`HC pE1
`2.2%
`
`1.4%
`1.5%
`2.4%
`
`2.5%
`2.9%
`5.2%
`
`Potency
`N/A
`
`108%
`101%
`20%
`
`Peptide mapping. A typical peptide mapping procedure was
`followed. Samples were diluted in reducing buffer (50 mM TRIS-
`HCl at pH 8.0, 6 M guanidine-HCl, 5 mM EDTA and 20 mM
`DTT) and were incubated at 56°C for 30 min. Each sample was
`cooled at room temperature for 5 min prior to alkylation with 50
`mM iodoacetamide at room temperature for 30 min in the dark.
`Each sample was then buffer-exchanged to digestion buffer con-
`taining 50 mM Tris buffer at pH 8.0, 2 M urea and treated with
`an aliquot of trypsin. The mixture was incubated at 37°C. The
`digestion was quenched by the addition of 20% trifluoroacetic
`acid (TFA). Agilent 6538 Q-TOF or Xevo G2 Q-TOF system
`mass spectrometers were set up with Agilent 1290 Infinity HPLC
`system or Waters Acquity UPLC. The samples were loaded on
`a C18 column held at 75°C. Mobile phase A was 0.05% TFA
`in water and mobile phase B was 0.05% TFA in acetonitrile.
`Gradient is adjusted to suit the needs of each molecule Peptide
`mapping analysis was performed by using Agilent MassHunter
`Qualitative Analysis software.
`Quantitation of mAb C deamidated peptide in plasma by
`multiple reaction monitoring. mAb C was spiked in rhesus
`monkey plasma and incubated for up to 21 days at 37°C. Blank
`plasma was prepared and incubated along with the spiked sam-
`ples serving as a negative control for the rhesus monkey plasma’s
`IgG2. mAb C was subsequently affinity-purified from the bio-
`matrices using a goat anti-human IgG-heavy and light chain
`monkey adsorbed antibody conjugated to an agarose based Affi-
`Gel hydrazine gel, by coupling the sugar residues of the antibody
`to the hydrazine groups on the gel. The eluates were digested
`by trypsin followed by endoproteinase V8 to obtain a reason-
`ably sized peptide amenable to mass spectrometry analysis. The
`digests were separated by reverse phase chromatography, using
`a very shallow gradient necessary to separate the peptides con-
`taining Asn33, isoAsp33 and Asp33. The separated peptides
`were analyzed online by multiple reaction monitoring (MRM)
`targeted analysis on a TSQ Vantage Thermo mass spectrometer.
`Areas under the chromatogram were subsequently extracted for
`each species using Skyline software (MacCoss lab, Department
`of Genome Sciences, UW) as a platform or Xcalibur (Thermo
`Scientific) and plotted against time.
`
`Potency by direct binding ELISA. Typical direct binding
`ELISA was used. The 96-well ELISA plates were coated with the
`appropriate antigen for 1 h at 25°C for mAb B or overnight at 4°C
`for mAb C. Plates were washed with TBS (mAb B) or PBS (mAb
`C) containing 0.05% Tween 20 and blocked with a blocking buf-
`fer for 1 h at 25°C. For the primary incubations, various con-
`centrations of mAbs were added into the pre-coated plates. The
`amount of bound mAb was measured by either alkaline phospha-
`tase (mAb B) or horseradish peroxidase (mAb C) labeled anti-
`human IgG. The enzymatic reaction was developed with either
`PhosphoGlo AP substrate (mAb B) or Pico SuperSignal (mAb
`C). The EC50 vales were determined by nonlinear regression.
`Surrogate solubility. While not strictly a maximum solubil-
`ity study determination, viscosity measurements in conjunction
`with protein concentration experiments in standard buffers are
`meant to evaluate the feasibility of 100–200 mg/mL solutions.
`Buffer exchanges were performed at 5°C in 50 kDa Amicon®
`Ultra – 15 centrifugal filter units, with sample to buffer volume
`ratio of 1:1 and three volume exchanges. mAb C concentration
`was determined by UV absorption at 280 nm using a calculated
`extinction coefficient.
`Viscosity. Viscosity measurements were performed at 30°C
`using mVROC instrument.
`Thermal melts (DSC). Differential scanning calorimetry
`(DSC) was performed with Microcal VP-DSC. 400 μL of each
`sample and matching buffer were placed in a 96-well plate, and
`the thermograms were acquired between 10 and 90°C at a scan
`rate of 90°C/h, with 5 min pre-scan equilibration. The data was
`processed by subtracting from each sample thermogram, a cor-
`responding placebo–placebo scan, fitting a bas