Journal of pharmaceutical sciences.
`v. 101, no. 8 (Aug. 2012)
`General Collection
`W1 J0829
`2012-08-21 09:19:03
`
`PROPERlY OF THE
`NATIONAL
`LIBRARY OF
`MEDICINE
`
`t,· /:if American Pharmacists Association·
`~rnprming medication u~e. Advancing patient care.
`
`APhA
`
`Ex. 2027-0001
`
`

`
`Journal of Pharmaceutical Sciences
`
`~
`~ A Publication of the American Pharmacists Association
`
`APhA
`
`fip
`
`A Publication of the Board of Pharmaceutical
`Sciences of the International Pharmaceutical
`Federation
`
`~ aaps
`
`American Assodation of
`Pharmaceutical Scientists
`
`Published in Cooperation wit_h the ~mer:ican
`Association of Pharmaceutical Sc1ent1sts
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, (Print ISSN:
`0022-3549; Online: 1520-6017), is published monthly on be(cid:173)
`half of the American Pharmacists Association by Wiley Sub(cid:173)
`scription Services, Inc., a Wiley Company, 111 River St., Hobo(cid:173)
`ken, NJ 07030-5774.
`
`Periodical postage paid at Hoboken, NJ, and additional offices.
`Postmaster: Send all address changes to the Journal of
`Pharmaceutical Sciences, Journal Customer Services, John
`Wiley & Sons Inc., 350 Main St., Malden, MA 02148-5020.
`Copyright and Photocopying Copyright© 2012 Wiley Pe-
`riodicals, Inc. and the American Pharmacists Association. All
`rights reserved. No part of this publication may be reproduced,
`stored or transmitted in any form or by any means without the pri-
`or permission in writing from the copyright holder. Authorization
`to photocopy items for internal and personal use is granted by
`the copyright holder for libraries and other users registered with
`their local Reproduction Rights Organisation (RRO), e.g. Copy-
`right Clearance Center (CCC), 222 Rosewood Drive, Danvers,
`MA 01923, USA ("http://www.copyright.com/"www.copyright.
`com), provided the appropriate fee is paid directly to the RRO.
`This consent does not extend to other kinds of copying such as
`copying for general distribution, for advertising or promotional
`purposes, for creating new collective works or for resale. Spe-
`cial requests should be addressed to: "mailto:permissionsuk@
`wiley.com"permissionsuk@wileycom
`Information for subscribers
`
`The Journal of Pharmaceutical Sciences is published in 12 is-
`sues per year. Institutional subscription prices for 2012 are: Print
`& Online: US$2139 (US), US$2196 (Rest of World), €1460
`(Eu-
`rope), £1155 (UK). Prices are exclusive of tax. Asia-Pacific GST,
`Canadian GST and European VAT will be applied at the appropri-
`ate rates. For more information on current tax rates, please go to
`"http://www.wileyonlinelibrarycom/tax-vat"www.wileyonlinelibrary.
`com/tax-vat. The price includes online access to the current and
`all online back files to January 1st 2008, where available. For other
`pricing options, including access information and terms and condi-
`tions, please visit "http://www.wileyonlinelibrarycom/access"www.
`wileyonlinelibrarycom/access.
`
`journal; the views and opinions expre~sed do not n~cessarily ~e­
`flect those of the Publisher, the Amencan Pharmactsts Assocra(cid:173)
`tion and Editors neither does the publication of advertisements
`constitute any e~dorsement by the Publisher, the Ameri~an Phar(cid:173)
`macists Association and Editors of the products advertrsed.
`Back
`issues: Single
`issues
`from current and
`recent
`volumes are available at the current single issue price from
`cs-journals@wiley.com. Earlier issues may be obtained from
`Periodicals Service Company, 11 Main Street, Germantown,
`NY 12526, USA. Tel: +1 518 537 4700, Fax: +1 518 537 5899,
`Email: psc@periodicals.com
`Delivery Terms and Legal Title: Prices inclu~e delivery
`of print journals to the recipient's addr~s.s. ~elrvery te~ms
`are Delivered Duty Unpaid (DDU); the recrprent rs responstble
`for paying any import duty or .tax.es. Legal title passes to the
`customer on dispatch by our drstnbu~ors.
`.
`.
`The contents of this journal are rndexed rn the followrng:
`Analytical
`Abstracts
`(RSC),
`Biological
`Abstracts®
`{Thomson
`lSI), BIOSIS Previews® {Thomson
`lSI), CAB
`Abstracts® (CABI), Cambridge Scientific Ab~tracts (CSA/
`CIG), CCR Database (Thom~on lSI), Chemrcal Abstracts
`Service/Sci Finder (ACS), Chemrstry Ser~er Compound Center
`(Thomson lSI), Chemistry Server Reactron Center (Thomson
`lSI), ChemPrep™ (Thomson lSI), ChemWeb (Chemlndustry.
`com), Chimica Database (Elsevier), Chromatography Ab-
`stracts (RSC) CSA Biological Sciences Database (CSA/CIG),
`CSA Environ~ental Sciences & Pollution Management Data-
`base (CSA/CIG), Current Chemical Reactions® (Thomson lSI),
`Current Contents®/Life Sciences (Thomson lSI), EMBASE/
`Excerpta Medica (Elsevier),
`Index Chemicus® {Thomson
`lSI), Index Medicus/MEDLINE/PubMed (N~M) •. !nternational
`Pharmaceutical Abstracts
`{Thomson Screntrftc), Journal
`Citation Reports/Science Edition (Thomson lSI), MDL Beilstein
`(Elsevier), Reaction Citation Index™ (Thomson lSI), Reference
`Update (Thomson lSI), Science Citation Index Expanded™
`(Thomson lSI), Science Citation Index® (Thomson lSI), .sCOPUS
`(Elsevier), SIIC Databases (SociedO:d
`lberoamencana de
`Informacion Cientifica), and Web of Scrence® {Thomson lSI).
`Production Editor: AI G. M. Liburd (email: jpsprod@wiley.
`com).
`
`Journal Customer Services: For ordering information, claims
`and any enquiry concerning your journal subscription please go to
`www.wileycustomerhelp.com/ask or contact your nearest office:
`Americas: Email: cs-journals@wiley.com; Tel: +1 781 388
`8598 or 1 800 835 6770 (Toll free in the USA & Canada).
`Europe, Middle East and Africa: Email: cs-journals@wiley.
`com; Tel: +44 (0) 1865 778315.
`Asia Pacific: Email: cs-journals@wileycom; Tel: +65 6511 8000.
`Japan: For Japanese speaking support, Email: cs-japan@wi-
`ley.com; Tel:+65 6511 8010 or Tel (toll-free): 005 316 50 480.
`Visit our Online Customer Get-Help available in 61anguages
`at www.wileycustomerhelp.com
`Disclaimer: The Publisher, the American Pharmacists Associa-
`Printed in the U.S.A. by the Sheridan Group
`This paper meets the requirements of ANSIINISO
`tion and Editors cannot be held responsible for errors or any con-
`sequences arising from the use of information contained in this .
`Z39.48-1992 (Permanence of Paper).
`Th1s matenal was~<J<pied
`
`View this journal online at wileyonlinelibrary.com/journal/jps
`Wiley's Corporate Citizenship initiative seeks to address
`the environmental, social, economic, and ethical challenges
`faced in our business and which are important to our diverse
`stakeholder groups. Since launching the initiative, we have
`focused on sharing our content with those in need, enhancing
`community philanthropy, reducing our carbon impact, cr~at!ng
`global guidelines and best practices for paper use, establrshrng
`a vendor code of ethics, and engaging our colleagues and
`other stakeholders in our efforts.
`Follow our progress at www.wiley.com/go/citizenship
`
`Ex. 2027-0002
`
`

`
`Native-State Solubility and Transfer Free Energy as Predictive
`Tools for Selecting Excipients to Include in Protein Formulation
`Development Studies
`
`DOUGLAS D. BANKS,1 RAMIL F. LATYPOV,1 RANDAL R. KETCHEM,2 JON WOODARD,1 JOANNA L. SCAVEZZE,1
`CHRISTINE C. SISKA,1 VLADIMIR I. RAZINKOV1
`
`1Department of Drug Product Development, Amgen Inc., Seattle, Washington 98119-3105
`
`2Department of Therapeutic Discovery, Amgen Inc., Seattle, Washington 98119-3105
`
`Received 29 March 2012; revised 8 May 2012; accepted 10 May 2012
`
`Published online 30 May 2012 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23219
`
`ABSTRACT: In the present report, two formulation strategies, based on different aggregation
`models, were compared for their ability to quickly predict which excipients (cosolutes) would
`minimize the aggregation rate of an immunoglobulin G1 monoclonal antibody (mAb-1) stored for
`long term at refrigerated and room temperatures. The first formulation strategy assumed that a
`conformational change to an aggregation-prone intermediate state was necessary to initiate the
`association process and the second formulation strategy assumed that protein self-association
`was instead controlled by the solubility of the native state. The results of these studies indicate
`that the stabilizing effect of excipients formulated at isotonic concentrations is derived from
`their ability to solubilize the native state, not by the increase of protein conformational stability
`induced by their presence. The degree the excipients solvate the native state was determined
`from the apparent transfer free energy of the native state from water into each of the excipients.
`These values for mAb-1 and two additional therapeutic antibodies correlated well to their long-
`term 4◦C and room temperature aggregation data and were calculated using only the literature
`values for the apparent transfer free energies of the amino acids into the various excipients and
`the three-dimensional models of the antibodies. © 2012 Wiley Periodicals, Inc. and the American
`Pharmacists Association J Pharm Sci 101:2720–2732, 2012
`Keywords: protein formulation; aggregation; conformational stability; solubility; thermody-
`namics; kinetics; transfer free energy; excipient; monoclonal antibody
`
`INTRODUCTION
`Protein biologics, and monoclonal antibodies (mAbs)
`in particular, are a fast growing modality in the phar-
`maceutical industry owing to their general greater
`in vivo predictability, specificity, and lower toxic-
`ity than small molecule pharmaceuticals.1–3 Unlike
`small molecule therapeutics, however, proteins are
`large complex molecules that are only marginally
`
`Abbreviations used: mAb, monoclonal antibody; IgG, im-
`munoglobulin G; Fab, fragment antigen binding; CH2, constant
`heavy chain domain two; CH3, constant heavy chain domain three;
`CD, circular dichroism; DSC, differential scanning calorimetry; Tm,
`melting temperature; SEC, size-exclusion chromatography; HMW,
`high molecular weight
`Additional Supporting Information may be found in the online
`version of this article. Supporting Information
`Correspondence to: Douglas D. Banks (E-mail: dbank@amgen
`.com)
`Journal of Pharmaceutical Sciences, Vol. 101, 2720–2732 (2012)
`© 2012 Wiley Periodicals, Inc. and the American Pharmacists Association
`
`thermodynamically stable and susceptible to numer-
`ous physical and chemical instabilities that might
`compromise their biological activity, or worse promote
`an immunogenic response.4,5 These attributes chal-
`lenge the protein scientist to deliver a formulation
`that is expected to maintain the protein therapeutics
`physicochemical properties over an extended period
`of time, usually ranging from 1 to 2 years at 2◦C–8◦C
`storage.6
`Exploring the formulation space to find the optimal
`solvent conditions for the long-term storage of pro-
`tein biologics can be a tedious and time-consuming
`process; therefore, formulation studies are typically
`made more efficient by employing a multivariant de-
`sign of experiment (DOE) approach wherein the effect
`of more than one formulation component on protein
`stability is studied at the same time.7 These stud-
`ies are best suited for testing continuous formula-
`tion parameters, such as excipient (cosolutes such as
`
`2720
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 8, AUGUST 2012
`
`Ex. 2027-0003
`
`

`
`NATIVE-STATE SOLUBILITY AND TRANSFER FREE ENERGY AS PREDICTIVE TOOLS
`
`2721
`
`sugars, polyols, amino acids, etc.) concentration and
`pH, where fractional-factorial experiments may be
`employed, but are less ideal for testing categorical
`parameters such as excipient or buffer type. In the
`latter case, full-factorial designs must be used to test
`each possible combination of formulation parameters
`on a given formulation response. This leads to large
`and cumbersome studies, if more than only a hand-
`ful of these discrete parameters are incorporated into
`a single study and exemplifies the need for smaller
`prescreen or predictive experiments that can be com-
`pleted within a reasonable timeframe to select only
`the top categorical formulation parameters to incor-
`porate into a DOE study.
`In the present report, two formulation strategies
`were tested for their ability to quickly predict which
`excipients would best minimize the association rate
`of an IgG1 mAb (mAb-1) after long-term storage at
`refrigerated and room temperatures by comparison
`with “real-time” data collected for nearly 1 year and 6
`months at 4◦C and 29◦C storage, respectively. Both
`of these strategies were based on different aggre-
`gation models; the first assumed that protein self-
`association is controlled by a partial unfolding event
`to some aggregation-prone intermediate state (the
`Lumry–Eyring model)8,9 and the second assumed
`that aggregation at low storage temperatures is in-
`stead controlled by the solubility of the native state.
`The first formulation strategy was tested by identify-
`ing an aggregation-competent intermediate state pop-
`ulated at high temperature and estimating its rate of
`formation at lower storage temperatures, as well as by
`empirically ranking the excipients effect on the mAb-
`1 thermal stability and aggregation rate at elevated
`storage temperatures. The solubility of the mAb-1 na-
`tive state in the presence of the different formulation
`excipients was tested in the second formulation strat-
`egy by ammonium sulfate precipitation.
`The results of these studies suggest that the rate
`of mAb-1 self-association is not controlled by con-
`formational stability under ideal storage conditions,
`but is instead influenced by the solubility of its na-
`tive state. It is believed that excipients affect pro-
`tein solubility because of their interactions with the
`solvent-exposed surfaces of the folded protein, which
`may be conveniently approximated using the appar-
`ent transfer free energies of the native state from
`water into each of the excipients.10 These values for
`mAb-1 and two additional mAbs correlated well to
`their long-term 4◦C and room temperature aggrega-
`tion data and were calculated using only the litera-
`ture values for the apparent transfer free energies of
`the amino acids into the various excipients and three-
`dimensional models of the antibodies. Under condi-
`tions wherein protein association and eventual for-
`mation of high-molecular-weight (HMW) aggregate
`are controlled by native-state solubility, such as phar-
`
`maceutically relevant 2◦C–8◦C storage temperatures
`at which proteins tend to be the most conformation-
`ally stable, we anticipate this methodology to prove
`useful as a practical means to select which excipients
`to include in formulation development studies.
`
`MATERIALS
`The IgG1 (mAb-1) and the IgG2s (mAbs 2 and 3) were
`recombinantly derived (Chinese hamster ovary cell
`expressed) and purified in house. Protein concentrati-
`ons were measured at 280 nm using extinction coeff-
`icients of 1.6, 1.8, and 1.4 mL/mg cm for mAb-1, mAb-2,
`and mAb-3, respectively. High purity ammonium
`sulfate, sucrose, sorbitol, L-proline, and PEG-6000
`were purchased from Sigma–Aldrich (St. Louis,
`Missouri). Sodium chloride,
`sodium phosphate
`(monobasic monohydrate and dibasic dihydrate),
`glacial acetic acid, and glycerol were purchased from
`J.T.Baker (Phillipsburg, New Jersey).
`
`METHODS
`Pharmaceutical Excipient Stability Studies
`Seven mAb-1 formulations were prepared in the fi-
`nal conditions of 100 mg/mL mAb buffered by 20 mM
`sodium acetate (pH 5.2). Five of these formula-
`tions contained 270 mM of the following excipients
`spiked in from 2 M stock solutions: sucrose, sor-
`bitol, L-proline, glycerol, and NaCl. The sixth formu-
`lation contained 2% (w/v) PEG-6000 [added from solid
`polyethylene glycol (PEG)] and the seventh formula-
`tion contained no excipient and was used as a con-
`trol. All seven mAb-1 formulations were filter ster-
`ilized in a laminar flow hood and 1.2 mL aliquots
`were dispensed into 3 mL borosilicate glass type 1
`vials and sealed with Daikyo rubber stoppers (Daikyo
`Seiko, Ltd., Tokyo, Japan). All formulations were then
`stored at 4◦C for 11.25 months and at 29◦C for 5.75
`months, during which time samples were pulled peri-
`odically and analyzed by size-exclusion chromatogra-
`phy (SEC).
`Each excipient was chosen on the basis of its differ-
`ent mode and degree of protein stabilization. Sucrose,
`sorbitol, and glycerol were chosen on the basis of the
`rational that they are preferentially excluded from
`the peptide backbone to varying extents and should,
`therefore, destabilize the unfolded state, thus mini-
`mizing the potential for partial unfolding of the na-
`tive state to an aggregation-prone intermediate.11–13
`PEG was chosen for the same reason, although it is be-
`lieved to be excluded from the unfolded state for steric
`reasons rather than any intrinsic unfavorable inter-
`actions with the peptide backbone.12,14 Proline was
`included because it has been shown to only weakly
`destabilize the unfolded state, but preferentially
`
`DOI 10.1002/jps
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 8, AUGUST 2012
`
`Ex. 2027-0004
`
`

`
`2722
`
`BANKS ET AL.
`
`interact with the exposed side chains of the native
`state, therefore, solubilizing it.12,15 Finally, NaCl was
`included in the study with the notion that low con-
`centrations of salt may either solubilize the native
`state (salt in) or enhance its conformational stability
`by shielding any local unfavorable electrostatic in-
`teractions. It became quickly apparent, however, that
`NaCl greatly increased the aggregation rate of mAb-1;
`therefore, this formulation was not considered beyond
`the 3-month time point.
`Stability studies for 100 mg/mL mAb-2 and 3 for-
`mulations were set up in an identical fashion as those
`described for mAb-1 and tested the excipients proline,
`sucrose, sorbitol, and glycerol for mAb-2 and proline,
`sorbitol, and glycerol for mAb-3. Both mAbs 2 and
`3 were stored at 4◦C and 25◦C during which time
`aliquots were pulled and analyzed by SEC. The stor-
`age times for mAbs 2 and 3 were 12 and 3 months,
`respectively.
`
`Pharmaceutical pH Stability Study
`The mAb-1 was formulated as a function of pH by dial-
`ysis into nine separate 20 mM buffer solutions rang-
`ing in pH from 4 to 8. Sodium acetate, succinate, and
`phosphate were used as the buffers between the pH
`ranges of 4–5, 5–6, and 6.5–8, respectively. Following
`dialysis, each solution was concentrated to 100 mg/
`mL, filter sterilized in a laminar flow hood and 1.2 mL
`aliquots were dispensed into 3 mL borosilicate glass
`type 1 vials and sealed with Daikyo rubber stoppers
`(Daikyo Seiko, Ltd.). Prior to being stored at 4◦C for
`3 months, the final pH of each formulation was mea-
`sured. This was necessary because the actual pH of
`the final formulations varied from the dialysis buffer
`pH, particularly at the lowest pH values because of
`the Donnan effect described elsewhere.16
`
`Size-Exclusion Chromatography
`Size-exclusion chromatography was carried out on
`an Agilent 1050 series quaternary pump LC system
`(Agilent, Palo Alto, California) equipped with a single
`TosoHaas TSK-gel SW3000xl 7.8 × 300 mm2 column.
`The column was equilibrated with mobile phase con-
`sisting of 100 mM sodium phosphate (pH 6.9) and
`150 mM NaCl. Fifty micrograms of antibody was in-
`jected onto the column and eluted isocratically at a
`flow rate of 0.5 mL/min with the mobile phase. Pro-
`tein elution was monitored at 230 nm.
`
`Circular Dichroism
`Near-ultraviolet circular dichroism (near-UV CD)
`spectra of mAb-1 were collected between 320 and
`250 nm using a bandwidth of 2 nm and a response
`time of 5 s on a Jasco J-815 CD spectrophotometer
`(Jasco, Tokyo, Japan) interfaced with a Jasco model
`PTC-423S/15 peltier type temperature-controlled cell
`holder (Jasco). A 5 mg/mL mAb-1 solution buffered at
`
`pH 5.2 with 10 mM sodium acetate was equilibrated
`in a 0.2 cm cuvette within the cell holder for 10 min at
`25◦C before recording the first near-UV CD spectrum.
`It was next heated to 67◦C and allowed to equilibrate
`for 10 min prior to recording the second spectrum. The
`final spectrum was recorded after cooling this same
`solution back to 25◦C and equilibrating for an addi-
`tional 10 min. Temperature-jump experiments were
`conducted by spiking 570 :L of buffer incubated at
`the desired final temperature (ranging from 63◦C to
`73◦C) in an Eppendorf ThermoStat plus temperature
`block (Eppendorf, Hamburg, Germany) into 30 :L of
`100 mg/mL mAb-1 stock kept at 25◦C in a similar
`temperature controller. This solution was then added
`to a 0.2 cm cuvette maintained at the desired final
`temperature in the instrument cell holder. Kinetic re-
`sponses were monitored at 254, 259, 264, and 269 nm
`using and bandwidth of 5 nm and a response time
`of 1 s.
`
`Differential Scanning Calorimetry
`The mAb-1 formulations used for the pharmaceutical
`stability study were diluted with their respective for-
`mulation buffers to 1 mg/mL and analyzed on a Micro-
`Cal (Northampton, Massachusetts) LLC VP capillary
`differential scanning calorimeter (DSC). Samples and
`buffer were heated from 20◦C to 110◦C at a scan rate
`of 150◦C/h. The filtering period was set to 4 s with
`medium feedback. The buffer was subtracted from
`each sample and the data normalized to convert to
`cal/(mol ◦C). These formulations were also analyzed
`using the same methods at an excipient concentration
`of 0.5 M.
`
`Accelerated Excipient Stability Study
`The aggregation kinetics of mAb-1 formulated in the
`identical conditions used for the long-term pharma-
`ceutical stability study were determined by incubat-
`ing samples at 63◦C in an Eppendorf ThermoStat plus
`temperature controller (Eppendorf) and quenching
`aliquots at increasing time points by placing at 4◦C.
`Samples were analyzed immediately using a Waters
`Acquity UPLC H-Class system (Waters, Milford, Mas-
`sachusetts) equipped with a BEH200 SEC 4.6 × 150
`mm2 column. Twenty micrograms of antibody was in-
`jected onto the column and eluted isocratically using
`a mobile phase consisting of 100 mM sodium phos-
`phate (pH 6.8) and 250 mM NaCl at a flow rate of
`0.2 mL/min; detection was at 280 nm. The total run
`time of this rapid SEC method was 6 min and the
`monomer eluted from the column after only 3 min.
`
`Ammonium Sulfate Precipitation
`Ammonium sulfate precipitation of each mAb-1 for-
`mulation was achieved by addition of ammonium sul-
`fate from a 4 M stock solution to final concentrations
`ranging from 1 to 1.72 M. The final solution conditions
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 8, AUGUST 2012
`
`DOI 10.1002/jps
`
`Ex. 2027-0005
`
`

`
`NATIVE-STATE SOLUBILITY AND TRANSFER FREE ENERGY AS PREDICTIVE TOOLS
`
`2723
`
`of each formulation were 5 mg/mL mAb-1, 270 mM ex-
`cipient (with the exception of the control and the 2%
`PEG formulation), 50 mM sodium acetate (pH 5.2) in a
`final volume of 400 :L. The 50 mM acetate concentra-
`tion was used to ensure that there was enough buffer-
`ing capacity to maintain the pH at 5.2 over the entire
`range of ammonium sulfate concentrations. Follow-
`ing ammonium sulfate addition, samples were mixed
`thoroughly by vortexing/inversion and then stored at
`29◦C and 4◦C for 3 and 24 h, respectively, in an Eppen-
`dorf ThermoStat plus temperature controller (Eppen-
`dorf). Following incubation, samples were centrifuged
`at 14,000g for 15 min in a centrifuge pre-equilibrated
`at the corresponding storage temperature. The super-
`natants were diluted with water and protein concen-
`trations were determined by absorbance at 280 nm.
`Care was taken to ensure that the 29◦C samples did
`not cool to room temperature prior to dilution.
`Ammonium sulfate precipitations of mAb-1 as a
`function of pH were conducted by spiking in ammo-
`nium sulfate to final conditions of 5 mg/mL mAb-1,
`1.27 M ammonium sulfate, 50 mM buffer (pH 5.2–7.1)
`in a final volume of 400 :L. Sodium acetate and suc-
`cinate were used to buffer these solutions at pH 5.2
`and 5.8, respectively, and sodium phosphate was used
`from pH 6.3 to 7.1. Solutions containing no mAb-1
`were prepared in the same manner to determine the
`actual pH of each solution because the addition of am-
`monium sulfate did decrease the pH, particularly at
`the higher pH values.
`
`Calculation of the Apparent Transfer Free Energies
`The methodology used to determine the apparent
`transfer free energy of the native state from wa-
`ter into a 1 M excipient solution ( Gtr,N) is well
`documented12,13,17 and Auton and Bolen18 have writ-
`ten an excellent review. Briefly, the chemical poten-
`tials of an amino acid at its solubility limit in water
`and in a 1 M excipient solution may be equated be-
`cause the chemical potential in the crystalline phase
`is invariant.19 The transfer free energy of the amino
`acid from water into the excipient solution ( Gtr) is
`then equal to the difference in the standard-state
`chemical potentials according to Eq. 1.20
`(cid:2)
`(cid:3)
`(cid:2)
`(cid:3)
`
` Gtr = :◦ = RT ln
`
`+ RT ln
`
`Sw
`Se
`
`(w
`(e
`
`(1)
`
`such values for the majority of the amino acids and
`model units of the peptide backbone (typically cyclic
`glycylglycine) have been determined in the presence
`of several excipients.11,17 Assuming that the apparent
` Gtr of all exposed groups of the folded protein are ad-
`ditive, the apparent Gtr,N from water to excipient is
`determined using them together in Eq. 2.13,22
`(cid:4)
` Gtr,N =
`
`ni"i gi
`
`(2)
`
`where ni is the number of an amino acid or backbone
`unit i of the protein, gi is the apparent transfer free
`energy for the side chain of amino acid i (or peptide
`backbone unit), and αi is the fraction of surface area of
`the amino acid side chain or backbone unit i exposed
`to solvent in the native state, determined from the
`protein’s three-dimensional structure or model. The
`full-length antibody models were built by first con-
`structing the fragment antigen binding (Fab) struc-
`tures using the Antibody Modeler tool in the Molec-
`ular Operating Environment (MOE) (Chemical Com-
`puting Group, Montreal, Quebec, Canada). The Fab
`structures were then used in conjunction with the
`structures 1HZH23 for IgG1 antibodies and 1IGT24
`for IgG2 antibodies to model the full-length antibod-
`ies using the Homology Model tool in MOE. Glycosy-
`lation, as was present in the full-length crystal struc-
`ture, was maintained, followed by covalent linkage
`to the heavy chain and isolated minimization. Per
`residue surface areas were calculated in the presence
`of all atoms using the Atom Surface Area function in
`MOE. Despite the aforementioned assumptions used
`to calculate the apparent native (or denatured) state
`apparent transfer free energies, thermodynamic prop-
`erties derived from them have agreed well with exper-
`iment, validating their predictive power.11
`
`RESULTS
`The Effect of Conformational Stability
`on mAb-1 Aggregation
`The effect of the seven cosolutes on the growth of
`mAb-1 dimer at 4◦C and 29◦C storage is shown in
`Figure 1. To test whether the differences in asso-
`ciation rates were the results of the different ex-
`cipients preventing the partial unfolding of the na-
`tive state to some aggregation-competent intermedi-
`ate state, we first tested the appropriateness of using
`the Lumry–Eyring model to describe the mAb-1 as-
`sociation reaction in the absence of excipient. This
`was carried out by monitoring the temperature de-
`pendence of the mAb-1 near-UV CD spectrum and
`assuming that dimerization and further aggregation
`would be controlled by its least-stable domain; an as-
`sumption made previously for the FC (fragment crys-
`tallizable) fusion protein abatacept.25 A change in the
`
`where S is the molar concentration of the amino acid
`at its solubility limit in water (Sw) and excipient (Se),
`and γw and γe are the corresponding activity coeffi-
`cients. The activity coefficients are typically ignored
`because of the difficulty in obtaining these values and
`based on estimates that have shown their ratio to usu-
`ally be close to one, particularly for the least-soluble
`amino acids.13,18,21 The Gtr values calculated in this
`manner are, therefore, “apparent” Gtr values and
`
`DOI 10.1002/jps
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 8, AUGUST 2012
`
`Ex. 2027-0006
`
`

`
`2724
`
`BANKS ET AL.
`
`Figure 1. The effect of excipient type on the rate of mAb-
`1 dimer formation for 100 mg/mL formulations stored for
`nearly a year at 4◦C (a) and for 5.75 months at 29◦C (b).
`The amount of dimer present was determined by SEC in
`the presence of 270 mM of the following excipients: NaCl
`), sucrose (
`), sorbitol (
`), glycerol (
`), and L-proline
`(
`(䉬). The rate of dimer formation was also determined in the
`) and in the presence of 2% (w/v)
`absence of excipient (
`PEG-6000 (䊊). All formulations were buffered at pH 5.2 by
`10 mM sodium acetate. Error bars represent the standard
`deviation of a mAb-1 standard injected multiple times and
`are a measure of the precision of the method. Lines are
`drawn to guide the eye.
`
`mAb-1 near-UV CD spectrum between 250 and 280
`nm first became apparent near 60◦C. Representative
`spectra of the same mAb-1 solution following sequen-
`tial 10 min incubations at 25◦C, 67◦C, and back to
`25◦C are shown in Figure 2. The nature of this ter-
`tiary structural change is unknown, but is likely due
`to the unfolding of the constant heavy-chain domain
`two (CH2) because it occurred in the vicinity of this
`domain’s melting temperature (supplemental data,
`Fig. 1) and was fully reversible; thermal denatura-
`
`tion of the mAb-1 Fab and constant heavy-chain do-
`main three (CH3) were both found to be irreversible
`processes. Despite the apparent reversibility of this
`structural change, the amount of dimer increased by
`approximately 2% after cooling mAb-1 back to 25◦C.
`If mAb-1 was allowed to incubate longer at 67◦C, the
`amount of dimer continued to increase. If the protein
`concentration was increased, the dimer appeared to
`reach a steady-state concentration where its rate of
`formation was offset by its rate of depletion to form
`HMW aggregate, even when incubated at the lower
`temperature of 63◦C (Fig. 2 inset). At these relatively
`high temperatures, the simplest aggregation mech-
`anism consistent with these data and the Lumry–
`Eyring model assumption is shown in Scheme 1. Here,
`the native state (N) unfolds to some partially de-
`natured intermediate (N∗) in a reversible reaction
`that is followed by the largely irreversible associa-
`∗) that over time
`tion reaction to nonnative dimer (N2
`is incorporated into HMW aggregate. To determine
`whether the initial mAb-1 dimerization reaction at
`4◦C and 29◦C is controlled by this same conforma-
`tional change, assumed to be rate limiting at low
`temperatures, the kinetics of this process was next
`followed by a series of temperature-jump experiments
`from 25◦C to final temperatures ranging from 61◦C to
`73◦C and monitoring the near-UV CD signals at 254,
`259, 264, and 269 nm. Kinetic responses at a given
`temperature for all wavelengths were fit globally to
`a first-order reversible reaction Eq. 3 using Igor Pro
`6.21 (WaveMetrics, Inc., Lake Oswego, Oregon).
`y(t) = y · exp(−kappt) + y∞
`
`(3)
`
`Here the total change in CD signal ( y) and the
`signal after equilibrium had been reached (y∞) were
`treated as free parameters and the apparent rate
`constant (kapp), which is equal to the sum of the
`forward (unfolding) and reverse (refolding) rate con-
`stants (kapp = kf + kr), was linked across all wave-
`lengths for a particular unfolding temperature. Rep-
`resentative kinetic traces, fit locally to Eq. 3 at 259
`nm, are shown in Figure 3a. The near-UV CD kinetics
`showed no protein concentration dependence, further
`supporting Scheme 1, but its highly nonlinear tem-
`perature dependence would question the use of this
`dimerization model at the lower formulation storage
`temperatures (Fig. 3b).
`This nonlinearity is likely a consequence of the re-
`versible nature of the two-state CH2 unfolding pro-
`cess. At high unfolding temperatures, kf can be ap-
`proximated by the kapp. As the unfolding temperature
`is decreased, however, the refolding rate constant (kr)
`
`Scheme 1.
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 8, AUGUST 2012
`
`DOI 10.1002/jps
`
`Ex. 2027-0007
`
`

`
`NATIVE-STATE SOLUBILITY AND TRANSFER FREE ENERGY AS PREDICTIVE TOOLS
`
`2725
`
`Figure 2. Near-UV CD spectrum of single 5 mg/mL mAb-1 sample following sequential
`10 min incubations at 25◦C (solid line), 67◦C (dotted line), and after return to 25◦C (dashed
`line). The sample was buffed at pH 5.2 by 10 mM sodium acetate in the absence of excipient.
`Inset: Despite the apparent reversibility of the CD spectrum, incubation of mAb-1 near the Tm
`of the CH2 resulted in an increase in the percent dimer ((cid:2)) that appears to reach a steady-
`state concentration where its rate of formation was balanced by its rate of depletion to form
`higher-molecular-weight aggregate (◦).
`
`is no longer insignificant and contributes more to kapp,
`resulting in its nonlinear temperature dependence.
`Below approximately 60◦C, the refolding dominates,
`making it highly unlikely that Scheme 1 describes the
`aggregation process