Journ al of
`v. 98
`pharmaceut·
`· no 4 (A
`1cal s ·
`Genelai. ,
`pr. 2009)
`Clences.
`VJ I IC) Col lection
`-
`- 829
`.
`LD09 OS Q1
`-
`: 03:46 57
`
`CSL EXHIBIT 1057
`CSL v. Shire
`
`Page 1 of 38
`
`

`

`Journal of Pharmaceutical Sciences
`
`~~
`~ A Publication of the American Pharmacists Association
`
`APhA
`
`.. t'~TIO.t;f
`
`~7&;,")\ A Publication of the Board of Pharmaceutical
`i
`R Sciences of the International Pharmaceutical
`~.,¥01J 11~''"
`Federation
`~ aaps
`
`American Association of
`Pharmaceutical Scientists
`
`Published in Cooperation with the American
`Association of Pharmaceutical Scientists
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, (Print ISSN:
`0022-3549; Online: 1520-60 17), is published monthly on behalf
`of the American Pharmacists Association by Wiley Subscrip(cid:173)
`tion Services, Inc., a Wiley Company, 111 River St., Hoboken,
`NJ 07030-577 4.
`Copyright © 2008 Wiley-Liss, Inc., a Wiley Company, and
`the American Pharmacists Association. All rights reserved. No
`part of this publication may be reproduced in any form or by
`any means, except as permitted under section 107 or 1 08 of
`the 1976 United States Copyright Act, without either the prior
`written permission of the publisher, or authorization through
`the Copyright Clearance Center, 222 Rosewood Drive, Dan(cid:173)
`vers, MA 01923, Tel.: (978) 750-8400, Fax: (978) 750-4470.
`Periodical Postage Paid at Hoboken, NJ and additional offices.
`The copyright notice appearing at the bottom of the first page
`of an article in this journal indicates the copyright holder's con(cid:173)
`sent that 2 copies may be made for personal or internal use, or
`for personal or internal use of specific clients, on the condition
`that the copier pay for copying beyond that permitted by law.
`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. Such
`permission requests and other permission inquiries should be
`addressed to the Permissions Department, c/o John Wiley &
`Sons, Inc., 111 River Street, Hoboken, NJ 07030; Tel.: (201)
`7 48-6011; Fax: (201) 7 48-6008; or visit http://www.wiley.com/
`go/permissions.
`Information for subscribers
`The Journal of Pharmaceutical Sciences is published in
`12 issues per year. Institutional subscription prices for 2009 are:
`Print & Online: US$1685 (US), US$1742 (Rest of World),
`€1123
`(Europe), £889 (UK). Prices are exclusive of tax.
`Australian GST, Canadian GST and European VAT will be
`applied at the appropriate rates. For more information on
`current tax rates, please go to www.wiley.com, click on Help
`and follow the link through to Journal subscriptions. The
`institutional price includes online access to the current and all
`online back files to January 1511997, where available. For other
`pricing options, including access information and terms and
`conditions, please visit www.interscience.wiley.com/journals
`
`Journal Customer Services: For ordering information, claims
`and any enquiry concerning your journal subscription please
`go to interscience.wiley.com/support 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@wiley.com; Tel: +65 6511
`8000.
`Delivery Terms and Legal Title: Prices include delivery
`of print journals to the recipient's address. Delivery terms
`are Delivered Duty Unpaid (DDU); the recipient is responsible
`for paying any import duty or taxes. Legal title passes to the
`customer on dispatch by our distributors.
`The contents of this journal are indexed in the following:
`Analytical
`Abstracts
`(RSC),
`Biological
`Abstracts®
`(Thomson
`lSI), BIOSIS Previews® (Thomson
`lSI), CAB
`Abstracts® (CABI), Cambridge Scientific Abstracts (CSAI
`CIG), CCR Database (Thomson lSI), Chemical Abstracts
`Service/Sci Finder (ACS), Chemistry Server Compound Center
`(Thomson lSI), Chemistry Server Reaction Center (Thomson
`lSI), ChemPrep™ (Thomson lSI), ChemWeb (Chemlndustry.
`com), Chimica Database (Elsevier), Chromatography Abstracts
`(RSC), CSA Biological Sciences Database (CSA/CIG), CSA
`Environmental Sciences & Pollution Management Database
`(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 (NLM), International
`Pharmaceutical Abstracts
`(Thomson Scientific), 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 (Sociedad
`lberoamericana de
`Informacion Cientifica), and Web of Science® (Thomson lSI).
`
`This paper meets the requirements of ANSI/NISO
`Z39.48-1992 (Permanence of Paper).
`
`This material was~o,pied
`at the N LM and may b<!
`5ubje~t US Copyright Laws
`
`Page 2 of 38
`
`

`

`Volume 98, Number 4, April 2009
`
`J ournal of
`Pharmaceutical
`Sciences
`= COMMENTARIES
`
`)
`
`Overlooking Subvisible Particles in Therapeutic Protein Products: Gaps That May
`Compromise Product Qual ity
`
`John F. Carpenter, * Theodore W. Randolph, Wim Jiskoot, Daan J.A. Crommelin,
`C. Russell Middaugh, Gerhard Winter, Ying-Xin Fan, Susan Kirshner, Daniela Verthelyi ,
`Steven Kozlowski, Kathleen A. Clouse, Patrick G. Swa nn , Amy Rosenberg, and Ba Ry Cherney
`Published online 14 August 2008
`
`1201
`
`/
`
`Biowaiver Monographs for Immediate Release Solid Oral Dosage Forms: Diclofenac
`Sodium and Diclofenac Potassium
`
`1206
`
`B. Chuasuwan, V. Binjesoh, J.E. Polli , H. Zhang, G. L. Am idon, H.E. Junginger, K.K. Midha,
`V.P. Shah, S. Stavchansky, J.B. Dressman, and D.M. Barends*
`Published online 27 August 2008
`
`~~· GLOBAL HEALTH COMMENTAR~
`
`Passing the Civilization Test
`
`Joseph T. Cunliffe Sr.
`Published online 24 November 2008
`
`REVIEWS
`
`)
`
`1220
`
`Effects of Glycosylation on the Stability of Protein Pharmaceuticals
`
`1223
`
`Ricardo J. Sola* and Kai Griebenow
`Published online 25 July 2008
`
`Principles, Approaches, and Challenges for Predicting Protein Aggregation Rates and
`Shelf Life
`
`1246
`
`Willi am F. Weiss IV, Teresa M. Young, and Christopher J. Roberts*
`Published online 6 August 2008
`
`Vo lume 98, Number 4 was mailed the week of March 23, 2009.
`
`In papers with more than one author, an asteri sk (*) in the byline indicates the
`author to whom inquiries shou ld be directed .
`atth e NLM and may b.e,
`This. mate·rial w as. copie-d
`Subject UCSCopyrigh.t Law s
`
`
`Page 3 of 38
`
`

`

`Vaccine Adjuvants: Current Challenges and Future Approaches
`
`1278
`
`Jennifer H. Wilson-Welder, Maria P. Torres, Matt J. Kipper, Surya K. Mallapragada,
`Michael J. Wannemuehler, and Balaji Narasimhan*
`Published online 14 August 2008
`
`From Natural Bone Grafts to Tissue Engineering Therapeutics: Brainstorming on
`Pharmaceutical Formulative Requirements and Challenges
`
`1317
`
`Biancamaria Baroli
`Published online 26 August 2008
`
`____ R_E_S_E_A_R_C_H __ A_R_T_I_C_L_Es _______ ~
`
`BIOTECHNOLOGY ..
`
`Structural Stability of Vault Particles
`
`-
`
`1376
`
`Reza Esfandiary, Valerie A. Kickhoefer, Leonard H. Rome, Sangeeta B. Joshi, and
`C. Russell Middaugh*
`Published online 6 August 2008
`
`Solid State Chemistry of Proteins IV. What is the Meaning of Thermal Denaturation
`in Freeze Dried Proteins?
`
`1387
`
`Michael J. Pikal, * Daniel Rigsbee, and Michael J. Akers
`Published online 14 August 2008
`
`Evaluation of the Effect of Non-B DNA Structures on Plasmid Integrity Via Accelerated
`Stability Studies
`
`1400
`
`S.C. Ribeiro, G.A. Monteiro, and D.M.F. Prazeres*
`Published online 8 September 2008
`
`Biochemical Mechanism of Acetaminophen (APAP) Induced Toxicity in Melanoma
`Cell Lines
`
`1409
`
`Nikhil M. Vad, Garret Yount, Dan Moore, Jon Weidanz, and Majid Y. Moridani*
`Published online 29 August 2008
`
`NMR Search for Polymorphic Phase Transformations in Chlorpropamide Form-A at High
`Pressures
`
`1426
`
`J. Wq_sicki, * D.P. Kozlenko, S.E. Pankov, P. Bilski , A. Pajzderska, B.C. Hancock, A. Medek,
`W. Nawrocik, and B. N. Savenko
`Published online 11 July 2008
`
`Synthesis, Characterization and In Vivo Activity of Salmon Calcitonin Coconjugated
`With Lipid and Polyethylene Glycol
`
`1438
`
`Weiqiang Cheng and Lee-Yong Lim *
`Published online 14 August 2008
`
`This m at .erial w asropie.d
`attne· NLM and may be
`Subject US Copyright Law s
`
`Page 4 of 38
`
`

`

`Single an? Double Emulsion Manufacturing Techniques of an Amphiphilic Drug in PLGA
`Nanopart1cles: Formulations of Mithramycin and Bioactivity
`
`Einat Cohen-Sela, Shay Teitlboim, Michael Chorny, Nickolay Koroukhov, Haim D. Danenberg,
`Jianchuan Gao, and Gershon Golomb*
`Published online 14 August 2008
`
`Spray-Dried Carrier-Free Dry Powder Tobramycin Formulations With Improved
`Dispersion Properties
`
`Gabrielle Pilcer, Francis Vanderbist, and Karim Amighi*
`Published online 27 August 2008
`
`1452
`
`1463
`
`PHARMACEUTIC~L TECHNOLOG)'
`
`·. :~
`
`..... -..-
`
`'" _ ~ ,.
`
`._ .. ,
`
`: .
`
`. ·. · .. ~~i~
`
`Pair Distribution Function X-Ray Analysis Explains Dissolution Characteristics of
`Felodipine Melt Extrusion Products
`
`K. Nollenberger,* A. Gryczke, Ch. Meier, J. Dressman, M.U. Schmidt, and
`s. Bruhne
`Published online 27 August 2008
`
`On-Line Monitoring of Pharmaceutical Production Processes Using Hidden
`Markov Model
`
`Hui Zhang, * Zhuangde Jiang, J.Y. Pi, H.K. Xu, and R. Du
`Published online 27 August 2008
`
`1476
`
`1487
`
`Mapping Amorphous Material on a Partially Crystalline Surface: Nanothermal Analysis
`for Simultaneous Characterisation and Imaging of Lactose Compacts
`
`1499
`
`Xuan Dai, Mike Reading,* and Duncan Q.M. Craig
`Published online 27 August 2008
`
`Hydrodynamic, Mass Transfer, and Dissolution Effects Induced by Tablet Location during
`D:ssolution Testing
`
`1511
`
`Ge Bai and Piero M. Armenante*
`Published online 9 September 2008
`
`PHARMACOKINETICS, PHARMACODYN.AMICS AND DRUG METABOLISM
`
`·
`
`Physiological Models Are Good Tools to Predict Rat Bioavailability of EF5154 Prodrugs
`from In Vitro Intestinal Parameters
`Masa hiro Nomot~,* Tomoko Tatebayashi, Jun Morita, Hisashi Suzuki, Kazumasa Aizawa,
`Tohru Kurosawa, and Izumi Komiya
`Published online 6 August 2008
`
`1532
`
`lnterindividual Pharmacokinetics Variability of the (X4J3 1 lntegrin Antagonist,
`4-[1 -[3-Ch loro-4-[N '-(2-methyl phenyl) u rei do] phenylacetyl]-( 4S)-fl uoro-(2S)-pyrrol id i ne-2-yl]
`methoxybenzoic Acid (D01-4582), in Beagles Is Associated with Albumin Genetic
`Polymorph isms
`
`1545
`
`Takashi Ito, * Masayuki Takahashi , Kenich i Sudo, and Yuichi Sug iyama
`Published online 14 August 2008
`
`This material w as copied
`at the· NLM and m ay b<e
`Su b<j e c:t US Copyright Laws
`
`Page 5 of 38
`
`

`

`Pharmacokinetic and Pharmacodynamic Evaluation of Site-Specific PEGylated
`Glucagon-Like Peptide-1 Analogs as Flexible Postprandial-Glucose Controllers
`
`1556
`
`Su Young Chae, Young Goo Chun, Seulki Lee, Cheng-Hao Jin, Eun Seong Lee, Kang Choon Lee, and
`Yu Seok Youn*
`Published online 14 August 2008
`
`Scintigraphic Study to Investigate the Effect of Food on a HPMC Modified Release
`Formulation of UK-294,315
`
`1568
`
`J. Davis,* J. Burton, A.L. Connor, R. Macrae, and I.R. Wilding
`Published online 27 August 2008
`
`A Cremophor-Free Formulation forTanespimycin (17-AAG) Using PEO-b-PDLLA Micelles:
`Characterization and Pharmacokinetics in Rats
`
`1577
`
`May P. Xiong, Jaime A. Yanez, GlenS. Kwon, Neal M. Davies, and M. Laird Forrest*
`Published online 27 August 2008
`
`Pharmacokinetics of Amitriptyline and One of Its Metabolites, Nortriptyline, in Rats:
`Little Contribution of Considerable Hepatic First-Pass Effect to Low Bioavailability of
`Amitriptyline Due to Great Intestinal First-Pass Effect
`
`1587
`
`Sao K. Bae, Kyung H. Yang, Dipendra K. Aryal, Yoon G. Kim, and Myung G. Lee*
`Published online 8 September 2008
`
`This mate ria I was H}pcied
`at the N LM and may b,e
`~ubject USCopcyright Laws
`
`Page 6 of 38
`
`

`

`Principles, Approaches, and Challenges for Predicting
`Protein Aggregation Rates and Shelf Life
`
`WilliAM F. WEISS IV, TERESA M. YOUNG, CHRISTOPHER). ROBERTS
`
`Department of Chemical Engineering, University of Delaware, 150 Academy St., Newark, Delaware 19716
`
`Received 17 April 2008; revised 11 June 2008; accepted 29 June 2008
`
`Published online 6 August 2008 in Wiley InterScience (www.interscience.wiley.com). DOl 10.1002/jps.21521
`
`ABSTRACT: Control and prevention of unwanted aggregation for therapeutic proteins
`is a ubiquitous hurdle during biopharmaceutical product manufacture, storage, ship(cid:173)
`ping, and administration. Methods to predict the relative or absolute rates of aggrega(cid:173)
`tion are therefore of great practical interest in biopharmaceutical research and
`development. Aggregation is often well-described as a multi-stage process involving
`unfolding or misfolding of free monomers, along with one or more assembly steps to form
`soluble or insoluble oligomers or higher-molecular-weight species. This report reviews
`the current state of the art in experimental and practical theoretical approaches that
`attempt to predict in vitro protein agb'Tegation rates or propensities relevant to phar(cid:173)
`maceutical proteins. Most available approaches fall within four primary categories. The
`principles and assumptions underlying each category are reviewed, along with advan(cid:173)
`tages and limitations in each case. The importance of appropriate experimental tech(cid:173)
`niques and models to probe and quantify the thermodynamics and/or dynamics of
`multiple steps or stages within the overall aggregation process is stressed. The primary
`focus is on aggregation in solution, relevant to parenteral dosage forms. Additional
`challenges are briefly reviewed. © 2008 Wiley-Liss, Inc. and the American Pharmacists
`Association J Pharm Sci 98:1246-1277, 2009
`Keywords: protein aggregation; biophysical models; physical stability; kinetics;
`protein formulation
`
`INTRODUCTION
`
`Aggregation in protein pharmaceutical products
`is typically undesirable from a number of per(cid:173)
`spectives. At a minimum, the resulting aggregates
`are an impurity that must be robustly controlled
`to meet regulatory constraints. In worst case
`scenarios, aggregates themselves may have safety
`implications, or impact pharmaceutical elegance
`
`Additional Supporting Information may be found in the
`online version of this article.
`Correspondence to: Christopher J. Roberts (Telephone: 302-
`831-0838; Fax: 302-831-1048; E-mail: cjr@udel.edu)
`Journal of Pharmaceutical Sciences, Vol. 98, 1246-1277 12009)
`'9 2008 Wiley-Liss, Inc. and the American Pharmacists Association
`
`and product marketability. 1-f> Control and/or
`
`prevention of aggregate formation during product
`manufacture, storage, shipping, and administra(cid:173)
`tion is therefore a common technical hurdle for
`successful biopharmaceutical development. 1
`•2•6•7
`In this context, the term aggregate refers to any
`self-associated state of a protein that is effectively
`irreversible under the conditions it forms, and
`often but not always is also a state in which
`biological activity of the constituent proteins is
`2
`8 Aggregates may be soluble or
`compromised. 1
`'
`'
`insoluble; the latter are often denoted as pre(cid:173)
`cipitates or particulates based on visual inspec(cid:173)
`tion, and typically jeopardize product quality or
`release. 1
`7 Soluble agf:,rregates can range from
`•
`dimers and small oligomers to so-called high
`
`1246
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 'Jil, NO. 4, APRIL 2009
`
`<+!WILEY
`lnterScience"
`
`This materialwasco-~ie<l
`at the NLM an<! may b.e
`Su b.je<t US Co~yright Laws
`
`Page 7 of 38
`
`

`

`PHEDICTING PROTEIN AGGREGATION HATES AND SHELF LIFE
`
`1247
`
`molecular weight (high-MW) aggregates com(cid:173)
`posed of tens, hundreds, or even more monomers
`per aggregate. 1·2·7- 11 High-MW aggregates are
`often considered ordered if they occur as long,
`rigid fibrils or filaments; typical of aggregates
`associated with a number of aggregation(cid:173)
`mediated neurodegenerative diseases or amyloi(cid:173)
`doses.8'12'13 By default, nonfibrillar high-MW
`aggregates are often termed amorphous. 10·14- 1G
`There is recent evidence that amorphous, high(cid:173)
`MW aggregates for pharmaceutically relevant
`systems such as G-CSF are well-described as
`linear, semi-flexible polymers. 10 An empirically
`common feature among aggregates from a variety
`of pharmaceutical systems is that the constituent
`monomers often show measurably elevated
`[)-sheet 2 ~ structure when compared to the folded
`•15 As such,
`or unfolded parent monomers. 1·2·8·1
`3
`;
`the term non-native aggregation is often applied
`to encapsulate the process of forming any or all
`of the different aggregate types summarized
`above. 2·17·18 This is also the convention adopted
`here. The term self-association is used to denote
`formation of small, soluble oligomers (native or
`non-native) that are reversible upon simple
`dilution with buffer. Protein crystallization and
`reversible, amorphous precipitation or salting-out
`processes that are commonly used to purify native
`proteins are denoted as native aggregation.
`Native aggregation is often governed practically
`by thermodynamics of phase separation, and is
`beyond the scope of this review. H\20
`Because non-native aggregation is typically
`irreversible for practical purposes, controlling
`aggregate formation inherently requires under(cid:173)
`standing and control of agf,'Tegation kinetics,
`although certain thermodynamic aspects are
`also of critical importance. 2•18 Fundamentals
`and practical analysis of experimental aggrega(cid:173)
`tion kinetics have been reviewed in detail else(cid:173)
`where.18·21 Only details relevant for the purposes
`of this report will be reviewed here. From the
`perspective of aggregation, as with many other
`degradation routes/·22 maximizing product shelf
`life is therefore equivalent to minimizing the rate
`of formation of (any) aggregate species or degra(cid:173)
`'2'7
`dants. 1
`The need for accurate and improved methods to
`predict aggregation rates or shelf life (tshclr) is
`driven in large part by the following disparity.
`Development timelines for innovator pharma(cid:173)
`ceutical products are often driven by a need to
`minimize the time from discovery-stage, lead(cid:173)
`candidate identification to clinical proof of con-
`
`cept. As such, the economic driving force is for
`multi-month or shorter timelines for bulk and
`clinical supply development and manufacture for
`a new candidate molecule. Practical considera(cid:173)
`tions also mandate commitment to a dosage
`form and formulation as early as possible during
`development. In contrast, regulatory constraints
`require multi-year, real-time stability testing to
`validate shelf-life expiry and product specifica(cid:173)
`tions. It is typically untenable for multi-year
`stability data at the target storage conditions to be
`available until the clinical program is in late
`stages. Simply put, the driver for improved shelf(cid:173)
`life prediction is one of minimizing risk: both
`the risk of costly time- and resource-consuming
`late-stage changes
`in product manufacture
`and formulation conditions, and the risk of
`unexpected and potentially hazardous aggregate
`formation in products upon prolonged storage or
`use. 5, 7,23,24
`Independent of shelf-life considerations, mini(cid:173)
`mization of aggregate formation during bulk
`active pharmaceutical ingredient (API) purifica(cid:173)
`tion and subsequent fill-finish operations is also
`critical from both product quality and regulat~ry
`considerations. 6 ·23 Currently, the primary dehv(cid:173)
`ery routes for therapeutic protein products are
`parenteral. 7·24- 26 Both lyophiles and ready-to-use
`liquid formulations are susceptible to aggregate
`formation during manufacture and storage .. For
`liquid formulations, protein mobility is re.labvel?'
`unrestricted and aggregation is often a pnmary If
`7
`not dominant degradation route. 1'2
`Even for
`'
`lyophiles, in which protein mobility in principle is
`severely restricted, aggregate formation fr?m
`solution is typically implicated during freezmg
`and/or reconstitution. 27- 29 In addition, the large
`majority of fundamental and/or systematic s~u­
`dies in the literature of protein aggregatiOn
`kinetics within and beyond pharmaceutical sys(cid:173)
`tems are available for bulk solutions. As such, the
`focus of this review is primarily upon prediction of
`protein aggregation in solution. This is no~ to
`downplay important and outstanding quest1~ns
`regarding approaches
`to predict aggregatiOn
`1
`"7 30 31
`t
`t' l
`during solid-state storage - '
`· or t 1e po en Ia
`l 28 ')9 3')
`'
`impact of adsorption at bulk interfaces, '
`·- ·· -
`but rather is dictated by practical limits of the
`scope of the article. Similarly, issues of cross(cid:173)
`aggregation with impurities (e.g., other proteins)
`are not reviewed here in the interest of space and
`because there is a relative dearth of reports that
`provide a tenable means to predict such issues,
`rather than simply documenting them.
`
`DOl 10.1 002/jps
`
`JOUI<NAL OF I'Hi\RMACEUTICAL SCIENCES, VOL. 9!3, NO.4, APRIL 2009
`
`This material wasm·pcied
`at the t~ LM and may b-e
`~u!J.ject US Copyright Laws
`
`Page 8 of 38
`
`

`

`1248 WEISS, YOUNG, AND ROBERTS
`
`In the next section, general aspects of non(cid:173)
`native aggregation are first reviewed as they
`pertain to aggregation kinetics and shelf life.
`This serves to highlight key thermodynamic and
`dynamic parameters that influence or control
`aggregation rates. A brief discussion is included of
`available experimental methods to quantify and
`monitor aggregation, along with recommenda(cid:173)
`tions based on strengths and limitations of each
`method. The subsequent section provides an
`overview of the four main categories of currently
`available and experimentally tested approaches
`for predicting aggregation shelf life of pharma(cid:173)
`ceutical or related systems. This includes a review
`of the basic principles and assumptions under(cid:173)
`lying each approach, as well as a discussion of
`their practical strengths and weaknesses. All of
`the approaches are shown to fall within a general
`framework that can be understood from the
`perspective of non-native aggregation as a
`multi-state process. A common problem for most
`practical prediction methods is shown to be that
`experimentally or
`theoretically
`convenient
`approaches often provide quantities that are only
`surrogates for the properties that fundamentally
`control aggregation rates, or they neglect impor(cid:173)
`tant contributions from one or more stages
`in aggregate formation. Also in this vein, the
`importance of directly measured aggregation
`kinetics is stressed, and a quantitative reevalua(cid:173)
`tion of Arrhenius versus non-Arrhenius shelf(cid:173)
`life extrapolation is presented. 'l'he next section
`briefly summarizes and highlights practical con(cid:173)
`siderations and concerns when selecting and
`implementing current or future approaches to
`predicting aggregation kinetics; the reader is
`referred to the final section for a summary list
`of nomenclature, abbreviations, and associated
`definitions used throughout this report.
`
`MECHANISM(S) AND KINETICS OF
`NON-NATIVE AGGREGATION
`
`Measurable or observed kinetics of aggregation
`depend on some but not all details of the
`mechanism of non-native aggregation. Specifi(cid:173)
`cally, they depend on the location of the rate(cid:173)
`limiting step or steps (RLS) in the mechanism. 18,21
`Any step that is "upstream" from a RLS is much
`faster than the RLS. Therefore, each upstream
`step reaches a local equilibrium, or it is said to pre(cid:173)
`equilibrate. This pre-equilibrium affects the con(cid:173)
`centration of species involved in the RLS, and
`
`JOURNAL OF I'HARMACEUTICAL SCIENCES, VOL. 9il, NO.4, APRIL 2009
`
`therefore also affects the observed rate for the
`overall process. 18' 3:l~:lG Steps that are "down(cid:173)
`stream" from the RLS are often considered to
`either be fast compared to the RLS, or to be
`kinetically unimportant. This second assumption
`is true only so long as later steps do not contribute
`in a kinetic sense to what is monitored experi(cid:173)
`mentally. This raises an equally important but
`often less well appreciated aspect of experimental
`aggregation studies. Qualitative and quantitative
`kinetics may be inadvertently influenced by which
`experimental assay is employed to monitor and
`quantify the aggregation process. With the above
`context in mind, the following subsections outline
`fundamentals of aggregation and measurement of
`aggregation kinetics that aid in understanding
`and evaluating different approaches to predicting
`agf,:rregation in later sections.
`
`Aggregation Kinetics and Shelf Life
`
`In each of the cases described below and in
`Figure 1, experimentally observed kinetics are
`often quantified in terms of the fraction of
`monomers converted to aggregate at a given time,
`m(t). For concreteness and for practical reasons, in
`this article m is defined as the mass fraction of protein
`that does not assay as aggregate in the experimental
`technique(s) being employed. Equivalently, 1-m is
`the fraction or the total protein mass that exists as
`aggregate at a given time (t), independent of aggregate
`size or solubility. These definitions are implicitly or
`explicitly assumed in most practical applications, and
`this is discussed further below.
`A convenient definition of shelf life (tsheir) that is
`often employed is to equate tshelf with the time for
`x percent loss of monomer (mass basis). Mathe(cid:173)
`matically, this is equivalent to
`
`m(tshelf) = ( 1 - 1~0)
`
`(1)
`
`In most pharmaceutical applications one is
`interested primarily in only small values of
`percent conversion to aggregate. Typical values
`of x are at most "'10 and often closer to 1 or
`lower. 7
`36 Kinetically, this means that one is
`'
`interested in "early time" or initial-rate kinetics.
`With the exception of kinetics that display a true
`lag time (see also the last section of this article),
`each case outlined in Figure 1 shows early-time
`kinetics that can be expressed as
`
`(2)
`
`DO/ I 0.1 002/jp'
`
`This materia I was co ~ied
`atthe N LM and may t>e
`~uhject US Copyright Laws
`
`Page 9 of 38
`
`

`

`kll
`F~
`(slow)
`
`A('()
`
`0. 1un F-
`(fast)
`
`(fast)
`
`PHEDICTING PHOTEIN AGGREGATION HATES AND SHELF LIFE
`
`1249
`
`Unfolding-Limited
`
`• soluble (large I small) agg.
`• insoluble agg. (ppt.)
`
`Association-Limited
`
`(slow)
`r.--k-.~,..,.L.,..,.' ) __ A._ ____ ,
`
`U ; ... %U2~ • i · ~- · small soluble oligomers
`~ ___ l _ _ _ _ _ _ _ _ _ _ _
`,... • high-MW,
`.
`.
`-
`soluble agg.
`• insoluble (ppt.)
`[If applicable]
`
`-
`
`-
`
`Nucleation I Growth I Condensation -Limited
`
`I
`X
`,. X
`k
`~c;cul
`~GC'J)
`23
`12
`U :;:=_ %U2~ • • ·~ , ux-.::. Ax- A 1-
`·
`\..
`/
`k
`x+ k
`V
`(fast)
`(slow)
`'-"'
`v
`• (slow); soluble oligomers
`• (fast); high-MW soluble agg.
`
`o
`
`• insoluble (ppt.)
`• high-MW agg.
`OR
`1 1 1
`t o
`T( )
`•• ·-A.-- .J. c
`k
`J
`
`a
`<•
`
`o
`" /
`
`Figure 1. Schematic representation of ag~:,rregation proceeding through a non-native
`monomer intermediate under native-favoring conditions. Unfolding is shown in a
`relatively simple form to indicate the limitations of commonly available experimental
`assays to distinguish beyond ensemble-averaged monomer conformational states. Dif(cid:173)
`ferent vertical levels in the diagram indicate the different stages of aggregation that
`must be accounted for in shelf-life prediction, depending on where the rate-limiting
`step(s) occur(s) for stable aggregate formation. Nomenclature is defined in the associated
`text of the Introduction Section.
`
`v is the apparent reaction order if one considers
`only monomer loss as a function of time at a single
`value of the initial protein concentration, and is
`typically equal to 1 or 2 (see also, below); hobs is the
`observed or effective rate coefficient for aggrega(cid:173)
`tion, and can be a complex function of a number
`of thermodynamic and kinetic parameters in
`Figure 1.
`Combining Eqs. (1) and (2) yields a relatively
`simple expression relating shelf life and hobs
`
`.-1)1/v
`X
`tshelf = 100 hobs
`(
`
`(3)
`
`In practice, one must set the value of x based on
`the product or specifications in question. As noted
`above, for pharmaceutical products an acceptable
`extent of aggregation is typically near or less than
`1% (i.e., x = 1). Eq. (3) is also a useful relation
`because much of the nonpharmaceutical litera(cid:173)
`ture on protein aggregation kinetics and aggrega-
`
`tion models characterizes aggregation rates in
`terms of hobs rather than tshclf· Based on Eq. (3),
`hobs and tshelf are used interchangeably through(cid:173)
`out this article. Similarly, if one quantifies
`aggregation rates in terms of the time for 50%
`monomer loss (t50), then x ~50 in Eq. (3).
`
`Overview of Non-Native Aggregation and Kinetics
`
`Figure 1 presents a schematic overview of non(cid:173)
`native aggregation in terms of multiple stages
`that can be important from the perspective of
`experimental kinetics. Different categories or
`classes of aggregation kinetics are shown hier(cid:173)
`archically, with the simplest (unfolding-limited)
`at the top and the most complex (nucleation/
`growth/condensation-limited) at the bottom.
`In general, the simpler categories are equi(cid:173)
`valent to simplifi.ed versions of the bottom-most
`case,
`in which rate-limiting steps occur at
`
`DOl I 0.1 OCJ2/jp'
`
`IOURNI\L OF 1'1-11\RM/\CEUTICAL SCIENCES, VOL. 98, NO. 4, APRIL 2009
`
`This materia I was <O·pi.ed
`at the NLM an<l may be
`
`Page 10 of 38
`
`

`

`1250 WEISS, YOUNG, AND ROBEHTS
`
`upstream positions such as unfolding or dimer(cid:173)
`ization.18
`The sequence of steps shown in Figure 1 is
`based on a starting point in which monomers are
`folded and typically in solvent conditions that
`favor the folded or native state (F). This assump(cid:173)
`tion is maintained throughout the article, as this
`is often the most relevant or desirable situation for
`pharmaceutical applications. There are at least
`five major stages or steps that can qualitatively
`or quantitatively affect observed aggregation
`kinetics (see also, Fig. 1).
`
`1. Folded monomers can unfold or partially
`unfold to one of a relatively large number
`of conformations that are collectively termed
`the unfolded monomer state (U). Additional,
`thermodynamically distinguishable confor(cid:173)
`mational states or intermediates may also
`occur for some proteins, but are not shown in
`Figure 1 for the sake of simplicity. As an
`example, if a stable folding intermediate (J u)
`that is more aggregation prone than U
`exists, one simply replaces U with I u in each
`of the schemes in Figure 1. Generalizing to
`more complex (un)folding schemes can also
`21 37-39 b t
`h
`.
`b
`d
`1s
`toregone
`ere as
`e
`one
`'
`u
`1"
`doing so does not significantly alter the dis(cid:173)
`cussions below. Unfolding from F is a unim(cid:173)
`olecular step with rate coefficient ku;
`refolding
`from
`the
`aggregation-prone
`unfolded state is also unimolecular, with
`rate coefficient kr. The standard free energy
`of unfolding (~G~n) between F and U is a
`function of temperature (T), pressure (p),
`and solvent composition. Unfolding and fold(cid:173)
`ing are in principle reversible for monomers,
`but are not necessarily at equilibrium during
`aggregation.
`2. Monomers may reversibly self associate
`to form soluble oligomers. In Figure 1 only
`self-association of U monomers is shown.
`This is based on the typical finding that some
`degree of unfolding is needed to facilitate
`association or later stages in aggregate for(cid:173)
`standard
`free
`mation.1'2'8'13'40-43 The
`energy change for the step u + ui of--t uit-1
`is denoted ~a}n1 , with i denoting the stoi(cid:173)
`chiometry of a' reversible oligomer, and the
`superscript (U) to be clear that it refers to
`association of U molecules. The forward and
`reverse rate coefficients for a given asso(cid:173)
`ciation step are also defined as shown in
`Figure 1.
`
`3. Soluble oligomers U; (i 2 2) may undergo
`additional conformational changes or inter(cid:173)
`nal structural rearrangement because their
`constituent U molecules are not tightly
`"bound" to one another, and because the U
`molecules retain some degree of conforma(cid:173)
`tional flexibility that free U monomers have.
`For simplicity, it typically is assum