`
`University Libraries
` University of Washington
`
`RAPID
`Lending — Article
`
`4/10/2015 8:15AM
`
`ILLiad TN: 1183176
`llllllll|ll|llllllllllllllllllll|lllllllllll|
`
`Borrower: RAPID:COD
`Shipping Address:
`University of Colorado at Boulder
`NEW: Main Library
`
`Odyssey: 206.107.44.246
`Phone:
`Email.
`Email:
`
`Needed By:
`Maximum Cost:
`EFTS?
`
`Patron:
`
`lLL Number: -9179297
`
`llllll|l||||l|l|l||||||||||l|||||||||||||ll||||||l
`
`Location: health serials
`Call #:
`title
`
`Journal Title: Critical reviews in
`therapeutic drug carrier systems
`
`Volume: 10 Issue: 4
`Month/Year: 1993
`
`Pages: 307-377
`
`Article Author: CLELAND, JL
`Article Title: THE DEVELOPMENT OF STABLE
`PROTEIN FORMULATIONS - A CLOSE LOOK AT
`PROTEIN AGGREGATION, DEAMIDATION, AND
`OXIDATION
`
`lSSN: 0743-4863
`OCLC #: 10588544
`
`Lending String:
`
`'" ”“633
`
`El Emailed Loc
`
`Special Instructions:
`
`Notes/Alternate Delivery:
`
`El Paged to 82
`
`El Paged to H8
`
`This article was supplied by:
`
`Interlibrary Loan and Document Delivery Services
`University of Washington Libraries
`Box 352900 - Seattle, WA 98195-2900
`
`(206) 543-1878 or Toll Free: (800) 324-5351
`OCLC: WAU - DOCLINE: WAUWAS
`interlib@u.washington.edu
`
`Odyssey
`
`AMGEN INC.
`Exhibit 1013
`
`
`
`EX.1013 - Page 1 Of 72
`
`
`Ex. 1013 - Page 1 of 72
`
`AMGEN INC.
`Exhibit 1013
`
`

`

`Critical Reviews in Therapeutic Drug Carrier Systems, 10(4):307—377 (1993)
`
`The Development of Stable Protein
`Formulations: A Close Look at
`Protein Aggregation, Deamidation,
`and Oxidation
`
`Jeffrey L. Ole/and, Michael F. Powell, * and Steven J. Shire
`
`Pharmaceutical Research and Development, Genentech. Inc.
`
`"' Address all correspondence to Dr. M. F. Powell, Genentech, Inc.. MS #10. 460 Pt. San Bruno Blvd, South San Francisco,
`CA 94080
`
`Releree: Dr. Paul Sleath, Immunex Corp., 51 University St, Seattle, WA 98101
`
`ABSTRACT: The biochemical literature has been surveyed to present an overview ofthe three most
`common protein degradation pathways: protein aggregation, deamidation. and oxidation. The mecha-
`nisms for each of these degradation routes are discussed with particular attention given to the effect
`of formulation conditions such as pH, ionic strength. temperature, and buffer composition. Strategies
`to reduce protein degradation are also discussed. These strategies are based on an understanding of
`the degradation mechanisms and the effect of changes in the storage conditions and formulation
`components on the degradation. The effects of each of the degradation routes on pharmaceutically
`relevant properties such as biological activity. metabolic half-life. and immunogenicity are summa-
`rized. Predicting a prior] the alteration of pharmaceutical properties caused by the three degradation
`routes is difficult, and must be determined on a case-by-case basis for each protein. The difficulty
`
`in predicting the effect of degradation and analyzing the temperature dependence of reaction rates in
`proteins results in longer development times for protein formulations than for small molecule
`formulations. Although the use of accelerated stability to predict protein shelf life is difficult,
`conditions are discussed whereby the Arrhenius equation can be used to shorten formulation devel-
`opment time.
`
`KEY WORDS: stability, degradation, lyophilization, parenteral, hydrolysis, precipitation, review,
`drug product. recombinant protein, immunogenicity, antigenicity, hydrogen peroxide, denaturation,
`solubility, half-life, metabolism, cosolvents, surfactants, residual moisture, freezing, catalysis, cyclic
`imide, bioactivity, conformation, excipients,
`ionic strength, mechanism. Arrhenius equation,
`temperature dependence.
`
`
`
`0743-4863/93/550
`
`© 1993 by CRC Press. Inc.
`
`307
`
`Ex. 1013 - Page 2 of 72
`
`Ex. 1013 - Page 2 of 72
`
`

`

`I. INTRODUCTION
`
`in designing a suitable drug formulation is to provide a
`The major goal
`compatible environment that ensures the maximum stability of the drug. This
`stability needs to be supported for a time period measured in years, in order to
`provide an adequate shelf life for the drug. For traditional, low molecular weight
`drugs, rational design of the formulation is well understood. It is based on an
`elucidation of the principal degradation routes and their kinetics, as well as an
`understanding of the mechanism of degradation. A significant body of literature
`exists in this field. For more complex protein and peptide drugs, no such literature
`is readily available on the pharmaceutical aspects of stability. The purpose of this
`review is to survey the biochemical literature in order to present an overview of the
`degradation mechanisms and major causes of instability in peptide and protein
`drugs. Some excellent reviews related to protein formulation have been published
`recently addressing the general stability of protein pharmaceuticals,‘ effect of
`formulation stabilizers,2 protein degradation pathways,3 and analytical methods to
`study protein degradation.4 The stability of protein pharmaceuticals has also been
`reviewed recently as part of an in-depth series on pharmaceutical biotechnology}6
`
`II. PROTEIN AGGREGATION AND FORMULATION STABILITY
`
`With the advent of biotechnology, greater amounts of protein have been made
`available to study various degradation reactions including aggregation. Early studies
`of proteins provided insight into the mechanisms of protein aggregation, but these
`studies failed to spark the interest of researchers in biotechnology.7 After the first
`successful expression of recombinant proteins, the formation of insoluble aggregates
`was observed in the host bacteria.8 These insoluble aggregates, referred to as inclu-
`sion bodies, reduced tne yield of soluble active protein and required dissociation to
`recover bioactivity of the expressed protein. For successful commercialization of
`recombinant proteins, methods were developed to recover native protein from intra-
`cellular protein aggregates (for reviews, see References 9 and 10). Many recent
`studies have been performed to determine the underlying mechanisms of both in vivo
`and in vitro protein aggregation (for reviews, see References 11 and 12).
`In contrast, there have been few detailed reviews of protein aggregation in
`formulations. The existing literature on protein formulations has usually focused
`on the chemical mechanisms for degradation and this literature is often directed
`toward research on therapeutic proteins. Usually, the therapeutic protein formula-
`tions described in the literature were designed to minimize aggregation since
`aggregation can have several detrimental effects including loss of activity, altered
`half-life, and increased immunogenicity. However, these studies are rarely de-
`signed to investigate the mechanisms of protein aggregation in the formulation. To
`minimize or eliminate aggregation, excipients such as sugars or surfactants are
`
`308
`
`Ex. 1013 - Page 3 of 72
`
`Ex. 1013 - Page 3 of 72
`
`

`

`oftei. empirically added to the formulation. A different mechanism of inhibiting
`aggregation may exist for each excipient. In addition, the physical state of the
`formulation (liquid or solid) affects protein stability. For solid formulation, the
`components of the formulation (such as buffer, excipients, etc), the method of
`achieving the dry state (such as lyophilization), and the conditions of the dry state
`affect protein stability. To assure protein stability for both liquid and solid formu-
`lations, the processes that cause protein aggregation must be understood.
`
`A. Mechanisms of Protein Aggregation
`
`To determine the factors influencing protein aggregation, the major causes of
`protein aggregation should be investigated. The distinction between association
`and aggregation must first be defined. Protein association is often considered a
`reversible process involving the interaction of two or more native protein mol-
`ecules and association may often result in reversible precipitation of the protein
`(see Section II.A.2). Protein aggregation usually involves the interaction of two or
`more denatured protein molecules and this process is often not reversible without
`dramatic changes in the solvent environment (see Section II.A.1). Therefore,
`protein denaturation must occur prior to the formation of aggregates according to
`this definition. Protein denaturation in solution is primarily caused by the solvent
`environment (pH, salt, cosolvents, etc.), temperature, or surface interactions. For
`lyophilized proteins, denaturation is dependent upon the final state of the protein
`as well as the changes in the protein solution during freeze-drying (see Section
`KID). The solvent environment may affect the physical state of the protein. The
`protein assumes the most thermodynamically favorable state in a given solvent
`environment. For example, in the presence of denaturants, the protein can assume
`a compact nonnative state with exposed hydrophobic residues that are normally
`buried in the protein core or it can form a random coil structure characteristic of
`a fully unfolded protein. When a protein is heated,
`it may also form different
`denatured states. At elevated temperatures, proteins often become more flexible
`and their collision frequency increases, resulting in a propensity to form aggre-
`gates. In addition, many surfaces used in pharmaceutical applications are hydro-
`phobic and protein adsorption to these surfaces can occur. Protein adsorption at
`hydrophobic surfaces may result in the accumulation of denatured protein at the
`surface, which results in an increased local concentration of denatured protein and
`subsequent aggregation. An air interface is a hydrophobic surface and, therefore,
`agitation of protein solutions with an air interface can also cause aggregation. The
`physical state of a protein and its stability depend on several environmental factors.
`Detrimental changes in the protein environment can result in protein denaturation.
`In general, the denatured state, which can be defined as any nonnative conforma-
`tion, often has exposed hydrophobic surfaces that may interact with other proteins
`both in solution and on surfaces.
`
`309
`
`Ex. 1013 - Page 4 of 72
`
`Ex. 1013 - Page 4 of 72
`
`

`

`1. Denaturation
`
`Many proteins are only marginally stable in solution; therefore, they are easily
`denatured and this results in aggregation. Nature has evolved methods to eliminate
`denatured proteins from the body and thereby reduce aggregation and prevent
`autoimmunity (see Section 11.3). These methods usually involve proteolytic deg-
`radation or glomerular filtration. In addition, intracellular processes remove dena-
`tured proteins and solubilize protein aggregates. One class of intracellular proteins,
`referred to as chaperonins, have the ability to bind denatured or aggregated proteins
`and facilitate their renaturation to the native state (for reviews, see References 13
`
`and 14). When the chaperonins become overwhelmed by the high expression level
`of a recombinant protein, intracellular aggregates (inclusion bodies) may form.”‘17
`Inclusion body formation of many recombinant therapeutic proteins (e.g., human
`insulin,8 human growth hormone (hGH),”‘-19 and human interferon-gamma”) re-
`quired extensive protein folding studies to recover the native protein. These
`studies, in turn, have provided a basic understanding of the processes of denatur-
`ation and aggregation.
`Two basic pathways have been observed for proteins during denaturation and
`folding. The simplest pathway involves a two-state model. In this model, the native
`state, N, exists in equilibrium with the denatured state, D:
`
`: D
`
`(1)
`
`The native state is not a rigid structure in solution and it can fluctuate to form
`several stable, active states that are difficult to detect by many methods.21 The
`denatured state is an ensemble average of several nonnative protein conformations.
`The characteristics and stability of the denatured state are dependent upon the
`environment (e.g., the solvent conditions). The presence of denaturants, cosolvents
`(see Section II.C), or extremes of pH can destabilize the native protein. As the
`protein denatures, the internal hydrophobic core residues are often exposed to the
`solvent.22 Interactions between hydrophobic residues in the interior of the protein
`determine the physical stability of the protein-"53-26 Stability of proteins has been
`successfully altered by making single residue substitutions in the hydrophobic
`core?"25 Thus, through protein engineering, it is possible to determine the critical
`hydrophobic residues and increase protein stability. Once the hydrophobic core
`residues are exposed to the solvent, they can interact with other hydrophobic
`surfaces such as container surfaces, air interfaces, or even with other denatured
`
`proteins. Exposure of these hydrophobic residues may proceed through a sequen—
`tial series of definable thermodynamic states referred to as intermediates. The
`formation of stable intermediate conformations that self—associate during folding
`has been observed for a number of proteinsflm'31 The second general pathway for
`denaturation and folding would then include these stable intermediate species:
`
`310
`
`Ex. 1013 - Page 5 of 72
`
`Ex. 1013 - Page 5 of 72
`
`

`

`,_~_
`AztIm
`
`(2)
`
`As shown in Equation 2, the native protein unfolds to form a native—like interme-
`diate, In, and this intermediate further unfolds to form other intermediates. Often,
`the protein will unfold to form an off-pathway or misfolded intermediate, 1m. These
`misfolded intermediates are often thermodynamically stable and their rate of
`folding to the native state can be significantly slower than the other intermediates.
`These misfolded intermediates usually contain nonnative disulfide bonds that must
`be reduced to allow folding to the native state.32 Misfolded intermediates and
`hydrophobic intermediates on the folding pathway may form aggregates, A. Al-
`though the misfolded intermediates and aggregates may eventually refold to the
`native state, the renaturation of these species is usually performed through the
`addition of denaturing agents such as sodium dodecyl sulfate (SDS), guanidine
`hydrochloride, or urea. If the misfolded intermediates or aggregates form nonna-
`tive disulfide bonds, the dissociation and folding to the native state require reduc-
`ing agents as well as a denaturant.
`Although many physicochemical interactions can lead to protein aggregation,
`several researchers have postulated that noncovalent
`interactions, specifically
`hydrophobic interactions, are the major cause of aggregation from the denatured
`state."-33‘37 Many studies of protein folding indicate that stable hydrophobic inter-
`mediates self—associate.27'3‘ For example, the hydrophobic molten globule interme-
`diates of carbonic anhydrase and bovine growth hormone self-associate during
`refolding.23'3°-33-39 If a thermodynamically stable intermediate state does not exist as
`suggested in the two-state model, hydrophobic interactions may still drive the
`aggregation of the protein since the gradual exposure of a hydrophobic surface will
`also increase the propensity of the protein to aggregate. For refolding of recombi—
`nant hGH, association of the protein was observed at high protein concentrations,
`but an intermediate species responsible for the aggregation was not identified.27
`Hydrophobic interactions are also the strongest nonpolar forces and may be the
`driving force for refolding.26 The endothermic l<inetics30-40-41 and loss of entropy“)-42
`observed during protein aggregation are further evidence of these nonpolar forces.
`Exposure of hydrophobic regions can occur during denaturation by chaotropic
`agent829-3“3140-“3 or by increased temperature.“"’*”'“49 Both denaturants and tempera-
`ture have been used to study the stability of proteins. If the two-state model as
`shown in Equation 1
`is assumed, the concentration of denaturant required to
`achieve an equal concentration of native and denatured protein will provide an
`indication of the thermodynamic stability of the protein. This approach can also be
`used to calculate the free energy of unfolding of the protein, which defines the
`thermodynamic stability of the protein.50 First, the fraction of denatured protein, fd,
`is calculated for a given condition (denaturant concentration, temperature, etc.) by
`
`311
`
`Ex. 1013 - Page 6 of 72
`
`
`
`Ex. 1013 - Page 6 of 72
`
`

`

`using a method that measures the extent of denaturation. Spectroscopic methods
`such as absorbance, circular dichroism, or fluorescence are commonly used for
`these studies. The fraction of denatured protein is then defined by the equilibrium
`between the denatured, D, and native, N, states:
`
`[D]
`KND
`f" =[N]+[D]:1+KND
`
`(3)
`
`where KND is the equilibrium constant. The equilibrium constant for denaturation
`is also related to the free energy of denaturation, AGOND, as a function of the
`denaturant concentration, [C]:
`
`KND =exp ——-—ND =exp ——————————([ D
`
`— AG°”2° + A c
`
`RT
`
`
`
`—AG°
`
`RT
`
`(4)
`
`where AGO ”:0 is the free energy of denaturation without denaturant and A corre-
`sponds to a linear relationship between denaturant concentration and change in free
`energy. By nonlinear least squares methods, experimental data can be used in
`Equations 3 and 4 with the denatured fraction as the dependent variable and the
`denaturant concentration as the independent variable. In addition, an equal concen-
`tration of both native and denatured protein is often achieved through heating to
`the appropriate temperature, usually referred to as the melt temperature, Tm. Both
`the heating and denaturation profiles of a protein are measured at equilibrium, and
`the results are used to calculate the thermodynamic stability of the protein as
`defined by Equations 3 and 4. These analyses are based on the assumptions that the
`protein does not form a stable intermediate state and does not aggregate.
`Another factor that can cause protein denaturation is surface interaction.
`Adsorption of proteins to interfaces has been extensively reviewed?”3 Adsorption
`studies have shown that proteins will adhere strongly to hydrophobic surfaces, but
`protein adsorption to hydrophilic surfaces is very weak.53 After adsorbing to a
`hydrophobic surface, a gradual denaturation of the protein may occur with the
`denatured protein remaining bound to the surface.”54 Once the buried hydropho-
`bic residues in the protein are exposed and adsorbed to the surface, dissociation of
`the protein from the surface can result in aggregation of the hydrophobic denatured
`protein. This process can be described by a series of reactions:
`
`N+S:NS:DS:D+S:A
`
`(5)
`
`where N is the native state, S is the surface, D is the denatured protein, and A is
`aggregate of denatured protein. Detailed models of protein adsorption for several
`
`312
`
`Ex. 1013 - Page 7 of 72
`
`Ex. 1013 - Page 7 of 72
`
`

`

`proteins have been described previously.53 Protein adsorption to surfaces can also
`decrease the thennostability of the protein?4 For example, denaturation and inac-
`tivation of B-galactosidase occurred upon interaction with wall surfaces at low
`temperatures and the inactivation was independent of the air—liquid interfacial
`area.55 Aggregation of denatured insulin on hydrophobic surfaces has been well
`studied. Monomeric insulin adsorbs to a hydrophobic surface or interface and is
`denatured?” Unlike many native proteins, native monomeric insulin has exposed
`hydrophobic surfaces that may initiate the surface adsorption. Subsequent denatur-
`ation and accumulation at the surface results in the formation of nonnative protein
`aggregates?“8 Surface adsorption that results in denaturation of the protein can be
`reduced or eliminated by either utilizing a more hydrophilic container or adding
`components to the formulation. Typically, glass vials are used for protein fonnu-
`lations, but silica, a slightly charged hydrophobic surface, can readily adsorb
`proteins. The use of more hydrophilic surfaces such as polyethylene oxide reduces
`protein adsorption and, thereby, prevents surface denaturation.53 Unfortunately.
`these materials are not widely used in protein pharmaceuticals. For therapeutic
`proteins, components are usually added to the protein formulation to prevent
`adsorption to the glass surface. These additives may inhibit surface adsorption by
`coating the surface and/0r binding to the protein. Serum albumin is often added to
`formulations because it competes for surface sites to prevent adsorption of the
`therapeutic protein. However,
`the addition of serum albumin may not be as
`effective as the addition of surfactants.“ Surfactants may bind to either the protein
`or the hydrophobic surfaces and inhibit protein adsorption. Copolymers of ethylene
`oxide and propyl oxide block copolymers called Poloxamers (e.g., Pluronics and
`Genapol series), and other polyethers such as the polysorbate series (e.g., Tween®)
`have been used to preVent protein adsorption. In particular, Genapol PF—IO,58
`Pluronic F-68,"0 Tween,57 and Triton X~10057 have been used to prevent surface
`adsorption and aggregation of insulin. Nonpolar solvents such as alcohols” and
`urea59 also provided a significant reduction in surface adsorption and denaturation
`of insulin. When adding surfactants and other stabilizers to the protein formulation,
`the effects of the additive on the pharmacological properties of the drug must be
`considered and additional toxicology studies may be required for materials that are
`not generally regarded as safe (GRAS). The additives may also cause leaching of
`the container components and stability studies must also address this possibility.60
`Proteins can also be denatured by other interfacial phenomena. Gas interfaces
`and shear forces from pumping or filtration can both cause significant denaturation
`of the protein. Like many solid surfaces, gas interfaces are nonpolar and can cause
`protein denaturation. The rate of denaturation at a gas-liquid interface is dependent
`upon the residence time of the protein at the interface.62 When shaking of protein
`solutions occurs in the presence of a gas-liquid interface, the rate of protein
`denaturation is increased resulting in the formation of both soluble and insoluble
`protein aggregates.63 Protein denaturation caused by agitation is the result of both
`the interfacial and shear forces on the protein. Shear denaturation of proteins can
`
`313
`
`Ex. 1013 - Page 8 of 72
`
`Ex. 1013 - Page 8 of 72
`
`

`

`occur as the result of agitation,“164 pumping,65v66 or filtration.64 As in the case of
`protein adsorption, surfactants have been added to formulations to reduce interfa-
`cial62 and shear-induced66 denaturation. The stability of many therapeutic proteins
`is often studied by subjecting the final configuration (e.g., stoppered vial) to
`shaking for several hours. If the protein in the final configuration retains its native
`state and does not aggregate, the formulation is considered stable against surface
`or shear—induced denaturation. These vigorous testing procedures are performed to
`assure that the patient is not adminiStered aggregated material as the result of
`storage or handling. Often, these formulations will require a stabilizing agent (e.g.,
`surfactant) to prevent protein denaturation.
`
`2. Charge Neutralization and Solubility
`
`Denaturation usually occurs prior to protein aggregation in solution, but the
`formation of native protein oligomers can also lead to aggregates of denatured
`protein. These phenomena often involve the association of the native protein as the
`result of changes in the solvent environment. The resulting protein oligomers may
`undergo subsequent denaturation over time as the result of the solvent conditions
`and protein-protein interactions. When a protein is placed in a high concentration
`of salt, the surface charges on the protein can become masked such that charge-
`charge repulsion does not occur. This surface charge neutralization can result in the
`association of neutral protein molecules. A similar phenomenon of charge neutral—
`ization can occur when the pH of the solution approaches the isoelectric point of
`the protein. As the isoelectric point is reached, the number of neutral protein
`molecules increases and association often occurs. Both salt and pH changes may
`lead to subtle changes in protein conformation in the a sociated state, resulting in
`the formation of irreversible aggregates. If the association process and salt or pH
`changes do not irreversibly alter the conformation of the protein, dissociation can
`occur by adjusting the pH or reducing the salt concentration. In the development
`of a stable formulation, the pH of the formulation is selected by taking into account
`factors such as the isoelectric point of the protein, solubility, and chemical degra-
`dation (e.g., deamidation). To reduce chemical degradation,
`the pH of many
`protein formulations is approximately 5 to 7. The isoelectric point of many hor-
`mones and blood proteins is also in the range of pH 5 to 7.67 Thus, the formulation
`pH must then be adjusted to avoid both chemical degradation and isoelectric
`precrpitation. To ensure adequate solubility, the protein formulation is usually
`maintained at least 0.5 pH units above or below the protein's isoelectric point.
`The solubility limit for native proteins in solution is dependent upon the
`solvent environment as well as the physical characteristics of the system (e.g.,
`temperature). Many therapeutic proteins can be maintained in a soluble state at
`concentrations (ug/mL to mg/mL) that are adequate for efficacy. Each protein must
`be studied for its solubility in a given solvent environment even though the
`
`314
`
`Ex. 1013 - Page 9 of 72
`
`Ex. 1013 - Page 9 of 72
`
`

`

`insoluble protein may not cause any adverse side effects such as increased immu—
`nogenicity (see Section II.B.2). Factors such as pl-I, excipicnts (see Section II.C),
`and buffer components inf“ Jence the protein solubility and the chemical and
`physical stability. These faC'ors also have an impact on protein lyophilization (see
`Section ILD). The protein in solution can occupy a specific volume of the solution
`as the result of its tertiary structure. The sum of the specific volumes of each
`protein molecule in solution will dictate the number of molecules that can be
`
`packed into a specific volume of water containing buffer components and excipi-
`ents. Therefore, as the protein concentration is increased, the packing density of the
`protein molecules increases, resulting in increased collisions and eventually the
`formation of a second phase (solid protein). This process usually occurs by a
`nucleation and growth mechanism where the critical nuclei are often soluble
`associated protein that rapidly grow to form insoluble protein precipitates or
`crystals.68 Several methods can be used to assess the solubility of a protein in a
`given solution. These methods often include adding solid protein to a solution and
`concentrating a solution by ultrafiltration. In most cases, the solubility limit of
`native proteins is not observed for therapeutic applications.
`
`B. Pharmaceutical Consequences of Protein Aggregation
`
`When developing stable formulations for a therapeutic protein, one must
`consider the effect of the formulation on the pharmacology of the drug. In particu-
`lar, if the formulation does not prevent denaturation and aggregation of the protein,
`the pharmacology, immunogenicity, and toxicology of the denatured or aggregated
`protein must be studied to determine its safety and efficacy. The Food and Drug
`Administration requires rigorous testing procedures to determine the stability of a
`pharmaceutical preparation over time. As the protein degrades chemically or
`physically, the activity, half-life, and immunogenicity of the protein can be altered.
`Therefore, several studies must be performed to determine the extent of degrada—
`tion that is acceptable for administration. The formation of soluble aggregates in
`a protein formulation can have a significant effect on the pharmacokinetics and
`immunogenicity of the protein. In addition, insoluble aggregates decrease the
`product quality and are generally considered unacceptable. The end user, physician
`or patient, is usually instructed not to administer a solution containing precipitates.
`An understanding of both the pharmacokinetics and immunogenicity of aggregated
`proteins is required for the development of a safe and efficacious product.
`
`1. Activity and Biological Half-Life
`
`Some proteins such as human insulin reversibly self-associate to form oligo-
`mers of the native protein. Recombinant human insulin was the first recombinant
`
`315
`
`EX.1013- Page 10 of 72
`
`Ex. 1013 - Page 10 of 72
`
`

`

`therapeutic protein approved for human use. Many of the early formulation stabil—
`ity studies were performed on recombinant human insulin. For insulin, it was
`critical to understand both association of native protein and aggregation of dena-
`tured protein, both for production from Escherichia coli69 and for development of
`a successful product formulation. Insulin exists in several associated states that can
`reversibly dissociate to the fully active monomer. The zinc hexamer of insulin is
`a complex of insulin and zinc that results in the formation of insulin crystals.57 This
`form of insulin has long-acting properties (e.g., extended biological half-life) when
`administered subcutaneously because it slowly dissociates into dimers and, even—
`tually, monomers that diffuse through the capillary membrane and into the blood.70
`Many recent studies have been performed to alter the self—association properties of
`insulin to produce stable dimeric or monomeric forms that are absorbed more
`rapidly and have a reduced biological half-life for treatment of hyperglycemia
`onset after meals“—73 These studies include engineering novel insulin mutants that
`do not associatem”2 and using surfactants to alter the distribution of associated and
`monomeric insulin.73 An understanding of insulin association has been required for
`successful treatment of diabetes. The study of insulin aggregation in solution has
`also been important for the development of insulin delivery devicesfié'm-M'75 Light
`scattering studies of insulin aggregation revealed the formation of large aggregates
`(100 to 200 nm).56'74'75 Insulin aggregates larger than the native hexamer are
`deleterious since the amount of active insulin is unknown. These aggregates may
`also invoke an unwanted and, perhaps, dangerous immune response.76
`
`2. Immunogenicity
`
`Before the dex ‘lopment of recombinant human insulin, insulin was obtained
`from animals (e.g., cattle and pigs). These animal-derived forms were immuno-
`genic even after extensive purification.77 However, the cause of this immunogenic-
`ity has never been clearly determined and it is likely that both the differences in
`the primary sequence of the human and animal-derived insulin and the integrity,
`degradation. and aggregation of the extracted animal forms invoked an immune
`response. Investigations of the relationship between aggregation and immune
`response have been performed with bovine serum albumin. These studies demon~
`strated that
`the administration of aggregated protein resulted in an increased
`immune response and the production of antibodies with a reduced affinity for the
`native protein.” Another recombinant therapeutic protein, hGH. has been studied
`for its immunogenicity as a function of aggregation. Initial therapies for hypopi-
`tuitary dwarfism utilized hGH extracted from the pituitaries of cadavers. The
`extracted protein often contained aggregated or degraded forms and these prepa—
`rations produced measurable amounts of antibodies to hGH after a single admin-
`istration. To assess the effect of hGH aggregation on immune response, a study was
`conducted with hGH in both the aggregated and monomeric forms. This study
`
`316
`
`EX.1013- Page 11 of 72
`
`Ex. 1013 - Page 11 of 72
`
`

`

`indicated that aggregates of hGH were the primary cause of immunogenicity in
`patients.79 Another study of hGH aggregation involved the administration of either
`the monomer or dimer of hGH. Both the dimer and monomer yielded equally low
`incidence of immunogenicity and the growth response was reported as equiva-
`lent.80 In vivo monomeric growth hormone binds and causes the dimerization of
`two growth hormone receptors.“ Therefore, the dimer of hGH must dissociate to
`have a biological effect. In addition, the presence of antibodies to hGH may not
`alter the effectiveness of the protein.82 These results suggest that immunogenicity.
`although undesirable, does not always diminish the efficacy of a protein therapeu-
`tic. The major concem in developing an immunogenic response to the therapeutic
`protein is the effect of this response on the patient’s own protein (cg. autoim-
`munity). Further research on the aggregation of other therapeutic proteins contin-
`ues to provide a better understanding of the role of aggregation in t