`Dosage Forms:
`Parenteral Medications
`Volume 1
`Second Edition, Revised and Expanded
`
`Edited by
`Kenneth E. Avis
`The University of Tennessee
`Memphis, Tennessee
`Herbert A. Lieberman
`HLH. Lieberman Associates,Inc.
`Consultant Services
`Livingston, New Jersey
`Leon Lachman
`Lachman Consultant Services
`Westbury, New York
`
`Marcel Dekker, Inc.
`
`New York « Basel « Hong Kong
`
`MAIA Exhibit 1028
`
`MAIA V. BRACCO
`
`IPR PETITION
`
`
`
`
`MAIA Exhibit 1028
`MAIA V. BRACCO
`IPR PETITION
`
`
`
`
`
`Library of Congress Cataloging - in - Publication Data
`
`Pharmaceutical dosage forms, parenteral medications / edited by
`Kenneth E. Avis, Herbert A. Lieberman, and Leon Lachman. -- 2nd ed.,
`rev. and expanded.
`p.-
`cm.
`Includes bibliographical references and index.
`ISBN 0-8247-8576-2 (v. 1: alk. paper)
`1. Parenteral solutions.
`2. Pharmaceutical technology.
`Kenneth E.
`II. Lieberman, Herbert A.
`III. Lachman, Leon.
`[DNLM: 1. Infusions, Parenteral.
`WB 354 P536]
`RS201.P37P48 1992
`615'. 19--dce20
`DNLM /DLC
`for Library of Congress
`
`I. Avis,
`
`2. Technology, Pharmaceutical.
`
`91 - 38063
`CIP
`
`This book is printed on acid-free paper.
`
`Copyright © 1992 by MARCEL DEKKER,INC. All Rights Reserved
`
`Neither this book nor any part may be reproduced or transmitted in any form
`or by any means, electronic or mechanical, including photocopying, micro-
`filming, and recording, or by any information storage and retrieval system,
`without permission in writing from the publisher.
`
`MARCEL DEKKER, INC.
`270 Madison Avenue, New York, New York 10016
`
`Current printing (last digit):
`0987654321
`
`PRINTED IN THE UNITED STATES OF AMERICA
`
`
`
`
`
`
`This material may be protected by Copyright law (Title 17 U.S. Code)
`
` ÿ
`ÿÿ
`ÿ
`ÿÿÿÿ
`
`7 P
`
`arenteral Products of Peptides
`and Proteins
`
`Yu-Chang John Wang
`California Biotechnology, Inc., Mountain View, California
`
`I.
`
`INTRODUCTION
`
`Scientific advances in molecular and cell biology have resulted in the develop-
`ment of two new biotechnologies. The first utilizes recombinant DNA to pro-
`duce protein products. Given the amino acid sequence of a protein, the gene
`responsible for producing such protein can be isolated, modified, and recom-
`bined with a plasmid DNA (Fig. 1). The modified plasmid can then be im-
`planted into a host cell. The recombinant DNA will be replicated and tran-
`seribed by the host cell to produce the specific protein in large quantities.
`Escherichia coli have been modified to produce growth hormone and growth
`factors, yeast to produce hepatitis B vaccine, and mammalian cells to produce
`erythropoietin. The advantages of such production are practically unlimited
`supplies of protein of high purity, and consistent quality in lot-to-lot produc-
`tion.
`The second technology involves fusion of antibody-producing B-lympho-
`cytes and myeloma (tumor) cells to form hybridomas (Fig. 2). Hybridomas
`grown in vivo in ascitic fluid or in vitro in tissue culture produce a single
`species of antibody, known as monoclonal antibody, which recognizes a spe-
`cific antigen.
`In contrast, conventional production methods in laboratory
`animals produce polyclonal antibodies that are comprised of a wide variety of
`antibodies with different affinity and antigenic specificities. Monoclonal anti-
`bodies, because of their unique properties of homogeneity, specificity, and
`affinity, can be used to:
`(1) isolate and purify proteins, e.g., Monoclat Fac-
`tor VIII:C is produced in this fashion; (2) neutralize a specific group of cells,
`e.g., Orthoclone OKT-3 neutralizes T-cells; (3) target™’specific cells or organs,
`e.g., antibodies to tumor antigens allow for tumor imaging or therapy by link-
`ing imaging radioactive isotopes or cytotoxic drugs to the antibodies.
`
`283
`
`
`
`
`
`
`284
`
`oe
`
`Donor DNA
`
`@
`
`®
`
`@
`
`@
`
`Wang
`
`—_
`
`/
`Restriction
`gs
`
`enzymes
`
`Ligases (
`Recombinant
`
`ONA
`
`
`
`Plasmid DNA
`
`Hos? call
`
`Figure 1 Recombinant DNA for protein production. The technique of recom-
`bining genes from one species with those of another. Plasmid is an extrachro-
`mosomal, independently replicating small circular DNA molecule.
`(1) Restric-
`tion enzymes cut DNA at specific places. The donor DNA, represented by a
`heavy band, contains the information to produce the protein of interest.
`(2)
`Ligase fuses donor DNA with plasmid together.
`(3) Calcium is used to open
`cell wall or cell membrane to allow recombinant DNA to enter the host cell (bac-
`terium, yeast or mammalian cell).
`(4) Replication of the host cells and recom-
`binant DNA increase its number. Expression of the DNA produces desired
`protein.
`
`New products approved by the FDA between 1982 and 1989 derived from
`these two new biotechnologies are listed in Table 1.
`Although these two new biotechnologies have greatly expanded the capa-
`bility for producing large amounts of high quality proteins, not all new pro-
`tein products are produced in this way. Small proteins and peptides, such
`as calcitonin, may be produced by chemical synthesis. Human serum albumin
`is sourced from human blood, urokinase from urine, and streptokinase from
`fungi. Most of these proteins and peptides are formulated as injectable prod-
`ucts, although there are some exceptions. Cyclosporin is given per os. Oxy-
`tocin and desmopresin are available as nasal preparations. A topical ointment,
`Elase, contains two lytic enzymes, fibrinolysin and desoxyribonuclease, in
`lyophilized form. Wound-healing growth factors are being developed for use
`as topicals.
`Among the parenteral formulations, most of them are in aqueous solution
`or in freeze-dried form for reconstitution. However, one of the potent luteiniz-
`ing hormone release hormone (LHRH) analogues, leuprolide, and bovine somato-
`tropin have been developed as erodable microspheres for depot injection.
`
`Il. CHARACTERISTICS OF PROTEINS AND PEPTIDES
`
`A. Protein Structures
`
`Proteins and peptides are made of amino acids linked by peptide bonds. The
`sequence of these amino acids defines the protein's structure. The sequence
`of amino acids in human insulin is illustrated in Figure 3. A protein may also
`show a secondary structure, which is formed by either intrachain or inter-
`chain hydrogen bonds. These bonds may result in either an a-helix or §-sheet
`
`284
`
`284
`
`
`
`
`
`
`Parenteral Products of Peptides and Preteins
`
`285
`
`Antigen
`
`Ascitic fluid
`
`* c=
`se
`S
`Oo.
`Oo
`
`B-lymphocytes 4\_Fusion select are ib wr
`
`Ce or
`pe %
`0%
`vr
`Myeloma
`OQ
`Hybridoma
`Geacat
`calls
`
`antibody
`
`Fermenter
`
`(1) A mouse or rat is immunized
`Figure 2 Monoclonal antibody production.
`with a specific antigen. The spleen is removed and antibody-producing cells
`(B-lymphocytes) are isolated.
`(2) Transformed myeloma cells from another
`animal are isolated.
`(3) B-lymphocytes are fused with myeloma cells deficient
`in a particular enzyme, to form hybridoma cells.
`(4) Only the fusion product,
`hybridoma cells, can grow indefinitely in culture. The unfused myeloma cells
`and lymphocytes die. The hybridoma cells are cloned and ones expressing
`the desired antibody are selected to culture in large amounts. The antibody
`can be produced by these hybridomacells in either ascitic fluid or fermenta-
`tion tanks. The ascitic fluid or culture media is then processed to yield puri-
`fied monoclonal antibodies.
`
`structure. Figure 4 shows a peptide chain of a protein coiled to form an a-
`helix.
`The tertiary structure or conformation of a protein refers to the spatial,
`three- dimensional structure of the polypeptide chain. Given the primary and
`secondary structure, there are four kinds of interactions cooperating which
`may contribute to the tertiary structure of a protein. They are:
`(1) hydro-
`gen bonds between residues in adjacent loops of the chain, (2) ionic interac-
`tions (salt bridges) between oppositely charged residues, (3) hydrophobic
`interactions between the aliphatic or aromatic residues, or (4) disulfide link-
`ages (Fig. 5). Protein, in its native state, exists as a tight, compact folded
`structure. The folded structure of insulin is shown in Figure 6. When the
`tertiary or secondary structure is destroyed, the protein unfolds and is in
`its denatured state.
`If dimers, trimers, or other oligomers are formed, the arrangement of
`these subunits is a quaternary structure. The best-known example of an
`oligomeric protein is hemoglobin that contains four chains and no disulfide
`linkages. The quaternary structure is essential to its oxygen-carrying activ-
`ity.
`Insulin also forms dimers and, in the presence of zinc, hexamers. The
`role of the zinc insulin hexamers is to provide a storage form that is thermo-
`dynamically stable and more resistant to enzymatic degradation than the un-
`associated monomer.
`
`285
`
`285
`
`
`
`
`
`
`286
`
`Wang
`
`Table 1 Biotech Products Approved in the United States Between 1982
`and 1989
`
`Company
`
`Date approved
`
`Indication
`
`Drug
`
`Insulin
`
`Human growth
`hormone® (hGH)
`
`Interferon
`A2b (IFN)
`
`Lilly
`
`Genentech
`Lilly
`
`Schering
`Roche
`
`October 1982
`
`October 1985
`March 1987
`
`June 1986
`June 1986
`
`June 1986
`
`Diabetes
`
`Dwarfism
`
`Hairy cell
`leukemia
`
`Reverse kidney
`rejection
`
`Vaccine against
`hepatitis B
`
`Thrombolytic
`agent
`
`Muromonab-CD3
`
`Ortho
`
`Hepatitis B
`vaccine
`
`Tissue plasminogen
`activator (t-PA)
`
`Merck
`
`July 1986
`
`Genentech
`
`October 1987
`
`Erythropoietin (EPO)
`
`Amgen
`
`June 1989
`
`Anemia
`
`Hepatitis B vaccine
`
`SmithKline
`Biologics
`
`August 1989
`
`Vaccine against
`hepatitis B
`
`“Growth hormone from Genentech differs from Lilly's product by a single
`methionine at the N-terminus.
`binterferons from Schering and Roche differ by one amino acid.
`
`In some proteins, there are carbohydrates attached to specific amino acid
`residues to form glycoproteins. For example, erythropoietin is a glycoprotein
`hormone with a molecular weight of 44 kilodaltons, having 40% of its weight
`accounted for by sialic acid, an amino sugar. Carbohydrate residues stabilize
`the tertiary structure and may influence receptor binding, bioactivity, and
`pharmacokinetics of the protein in vivo.
`
`B.
`
`Isoelectric Point
`
`In the course of a potentiometric titration of a protein there is a pH at which
`the mean charge on the protein is zero, This pH is termed the isoelectric
`point, pl. Table 2 provides examples of pIs for a number of therapeutic pro-
`teins and peptides. The pI may be approximately calculated from the amino
`acid composition data, i.e., pl = (pK; + pKg + pKg.
`.
`. + pK,,)/n for n ioniz-
`able groups. However, because the dielectric constant in the immediate vicin-
`ity of an ionizable group depends on protein structure and, because hydrogen
`bonding may alter dissociation constants (Kgs), the true pI can differ signifi-
`cantly from the calculated one.
`Just like pKgs of small organic compounds, the pI of a protein plays an
`important role in solubility.
`In general, near the pH of the isoelectric point,
`protein solubility is at its minimum. This trend is similar to the low solubility
`exhibited by a zwitterion compound.
`
`286
`
`286
`
`
`
`
`
`
`Parenteral Products of Peptides and Proteins
`
`287
`
`s—-s
`
`
`pooozys(S&1oraGLIAL24d4dATOBay|nqo[<t5shoTeA|NeTIALeTETYOTS|TEANOTSTHzegTDShQNe]STHUTO|USYTPAaqa°aN
`OfSt0zSTorST
`
`
`
`moo@Joseazoued94}Worpe}B[OSTSUT[NSUTysnoyyly‘eseyxUTepyMsrpTeusezuyueseysuTeyoay}jo
`
`
`
`
`
`ey}pue[TZUMUTeYydVYey}‘SUTeYyoOM}seyUTMsSUL“UPNSUTuewnyjosduenbespoeoulwy¢eanbi4
`
`
`
`
`auopuB‘SasBYUT[-SSOAOapIJ[NSIpOM}AqpeyoeuuodaresureyoOM}CYL“SpeOUTUTEOFUFIMureyo
`
`
`
`
`uedUT[NSUTUBUINY,=*“pexoqoie‘satoedsSuourejuelseaul‘*e*t‘senp{sedploeourwepaaresuooATySTYy
`
`
`
`
`
`
`-suei}orjewAzueueAqUTNsUryaodwoay10VNJUeUIqWODeTYIMyseak10909"gyAQpeonpordaq
`
`
`auL‘UT[NSUTUBUINYYIM[eoT}UeprjoueBAou}‘sjuerjedotjeqerpjoJUSUI}ee1}UTpesnodB31d@20
`
`
`
`
`
`
`
`
`ay}Sutwaoysny}‘O&goutueyeay}LoyeuTUOeTY}BseinyisqnsApeaTOeTesYOTUMsseoo0rduorjepyded
`
`
`[wooousyS49241)ATT795eento[Tea®TT4T9¢HNusyngnayUTDryney19s[s49|
`/ss/|IzscIoTs¢
`
`
`
`
`
`
`
`*sernoe[owuT;NsuUT-uBeUNY
`
`287
`
`287
`
`
`
`
`
`
`
`288
`
`Wang
`
`Figure 4 Peptide chain of a protein coiled to form an a-helix. Configuration
`of the helix is maintained by hydrogen bonds shown as vertical dotted lines.
`
`C. Degradation Through Covalent Bonds
`
`The degradation of proteins and peptides can be divided into two main cate-
`gories:
`those that involve a covalent bond and those involving a conforma-
`tional change. The latter process is often referred to as denaturation.
`Numerous types of chemical reactions leading to modification of the co-
`valent structure are possible in a protein. The following describes only those
`that are likely to be seen in a protein or peptide formulation.
`
`Peptide Fragmentation
`
`The peptide bond (RNH—CO—R) is much moreresistant to hydrolysis than is
`the ester linkage (R—O—CO—R) and peptide bonds are considered stable un-
`
`288
`
`288
`
`
`
`
`
`
`Parenteral Products of Peptides and Proteins
`
`289
`
`Disulfide
`Bond
`
`
`
`
`fonic
`lnteraction
`
`Figure 5 Interactions that stabilize folded conformation.
`
`less hydrolysis is assisted by a neighboring group. The formulation factor
`that most influences the hydrolytic rate is solution pH. The rate of hydroly-
`sis is in direct proportion to the activity of hydronium, or hydroxide ions,
`when in acidic or alkaline pHs, respectively. A rate-pH profile for hydrolysis
`of tetraglycine is shown in Figure 7. The reaction minimum is in the region of
`PH 3to 5 [2]. Published examples of stability studies on therapeutic peptides
`are: degradation of nafarelin [3], thermal stability and degradation mecha-
`nism of secretin [4], acid hydrolysis of captopril, a quasidipeptide [5], and
`heat stability of urokinase [6].
`Many peptides are stable enough to formulate as a ready-to-use parenteral
`formulation. For instance, D-Trp-LHRH was stable at 60°C for 5 days and
`seven cycles of freezing and thawing; gramicidin, although used only for derma-
`tological indications when dissolved in glycols, can be autoclaved.
`Information
`available on package inserts also shows it is possible to have good long-term
`stability of peptides. Leuprolide injection has a shelf-life of at least 2 years
`under refrigeration and 3 months at room temperature. LH-RH injection (Rele-
`fact) does not even require storage under refrigeration. Oxytocin injection
`was reported to be stable at room temperature for 5 years [7]. Protirelin, a
`tripeptide (PyrGlu-His-Pro), is stable for 20 hr at 80°C at both pH 3.3 and
`pH 6 [8].
`Certain amino acids form the weak link of the chain. The bond between
`aspartic acid and proline or tyrosine is sensitive to acid hydrolysis [9]. The
`resultant products are peptides with aspartic acid at the C-terminus. The C-
`terminal peptide bond adjacent to serine is also a reactive one due to the neigh-
`boring group effect of the alcohol on serine [10].
`
`289
`
`289
`
`
`
`
`
`
`290
`
`Wang
`
`Figure 6 Folded three-dimensional structure of insulin. The A-chain is ar-
`ranged between the two terminal arms of the B-chain and lies on top of the
`central a-helix (B9-Bj9). The hydrophobic regions are responsible for the
`association of monomer into dimers and hexamers.
`(Adapted from Ref. 1.)
`
`Table 2 pls of Selected Proteins
`
`
`
` Proteins pI
`
`Pepsin
`
`Erythropoietin
`
`Serum albumin
`
`Growth hormone
`
`Insulin
`
`Interleukin- 2
`
`Calcitonin (salmon)
`
`Chymotrypsin
`
`Basie fibroblast growth factor
`
`Interferon- y
`
`Lysozyme
`
`Nafarelin
`
`1.0
`
`3.5-4.0
`
`4.8
`
`5.0
`
`5.4
`
`7.8
`
`7.8
`
`9.1
`
`9,8
`
`10.0
`
`11.0
`
`11.3
`
`
`
`Protamines 12.0
`
`290
`
`290
`
`
`
`
`
`
`Parenteral Products of Peptides and Proteins
`
`291
`
`=2
`
`
`
`LogKopa(hr”)
`
`pH
`
`Figure 7 Rate-pH profile for the hydrolysis of tetraglycine. Peptide solu-
`tions containing phosphate buffer, 0.03 M, and sodium chloride, 0.1M, were
`incubated at 60°C. Samples were analyzed by HPLC assay. At pHs 1 to 3,
`the slope is -1, indicating an acid catalyzed reaction.
`In the pH region 4 to
`7, the reaction is hydroxide ion catalyzed. The plateau at pH 8 is caused by
`change in ionization state, zwitter ion to negatively charged ion, which is less
`susceptible to hydroxide-ion attack [2].
`
`Deamidation
`
`Deamidation refers to the removal of ammonia from the amide (RCONHg) moiety,
`whetherit is at the end of a C-terminus or at asparagine (Asn) or glutamine
`(Gln) residues. Stability studies of insulin [11], ACTH [12], human growth
`hormone [13], and RS-26306 [10], an LH-RH analogue, have shown this route
`to be the major factor resulting in instability of these proteins and peptides
`in aqueous solution.
`In acidic media, the peptides deamidate by direct hydrolysis. Thus, an
`Asn or Gln residue yields an Asp or Glu peptide, respectively. The neighbor-
`ing amino acid does not affect the deamidation rate. Hydrolytic mechanisms
`in neutral or alkaline pHs are more complex, however. Under these condi-
`tions, the side chain carbonyl group on the Asn or Gln residue attaches to
`the nitrogen atom on the peptide backbone to form a cyclic imide intermediate.
`Depending on which bond in the cyclic imide breaks (Fig. 8), the reaction
`product can be:
`(1) the des-amido peptide, (2) the isopeptide, or (3) D-
`isomers. The formation of isopeptide is sometimes referred to as transpeptida-
`tion because an extra methylene group is inserted to the peptide backbone.
`At neutral to alkaline pHs, the rate of deamidation is significantly affected by
`the size of the amino acid on the C-terminal side of the Asn or Gin residue.
`In general, Asn is more labile than Gln, and is most labile when adjacent to
`glycine, which is the least obstructive to the formation of a cyclic imide.
`Deamidation rate-pH profile is V-shaped [14], usually with a minimal rate
`at a pH of about 4 to 5 (Fig. 9).
`In a number of synthetic peptides the half-
`
`291
`
`291
`
`
`
`
`
`
`292
`
`Wang
`
`1
`-C-O° RO
`CHa
`|
`Jirarte
`|
`R'-NH-C—<”
` Deamidation
`§ for Asn & Gin
`
`0
`
`0
`
`°
`
`0
`
`=
`
`”
`a” eo
`1
`ott
`R'-NH—-CH-C-NH-CH-C ~
`i
`NH,
`a i
`a
`CHa yy Yo
`.
`Tl
`fe
`Che
`|
`N—CH—C~ aga
`NH-CH-C ~
`R'-NH-C__*b
`R'-NH C~¢ 09
`Hc §
`6 Transpeptidation
`oO
`
`a
`
`c
`
`Imide Formation
`
`oO
`R
`me
`&
`ey T.
`UH N-CH-C~
`R'=NH-CW °7 Fracemization
`
`il
`°
`
`Figure 8 Reaction pathways in deamidation of asparagine. Deamidation of
`asparagine or glutamine goes through an intermediate, a cyclic imide. Three
`routes of degradation can take place, which leads to products of deamidation
`(removal of the amino group), transpeptidation (insertion of an extra methy-
`lene group), and racemization (D aspartic or glutamic acids).
`
`Logky(hr)
`
`ro
`
`0
`
`2
`
`4
`
`8
`
`10
`
`12
`
`6
`pH
`
`Figure 9 Deamidation rate-pH profile of an active segment of adrenocortico-
`tropic hormone, Val-Tyr-Pro-Asn-Gly-Ala, in aqueous solution at 37°C and
`ionic strength of 0.5. The apparent rate is followed by the disappearance of
`the parent peak in HPLC.
`(From Ref. 14.)
`
`292
`
`292
`
`
`
`
`
`
`Parenteral Products of Peptides and Proteins
`
`293
`
`lives of deamidation reactions of Asns at 37°C in pH 7.2 phosphate buffer
`range from 2 days to nine years [15]. Not all Asns are equally labile; those
`buried within the interior portion of a protein are inaccessible to water and
`thus less reactive.
`In a large protein, secondary and tertiary structures
`play an important role in determining the site and the rate of deamidation.
`There are three asparagines in insulin (Fig. 3), the monodesamido-(A21)-
`insulin is the prevailing degradant formed in acid solution.
`In neutral solu-
`tion, deamidation is slow at 5°C. However, when accelerated at higher stor-
`age temperatures, deamidation of Asn at the B3 position can be detected [11].
`Figure 10 shows deamidation of insulin in neutral formulations at three differ-
`ent temperatures. With growth hormone, there are nine asparagines among
`191 amino acids, and deamidation oceurs primarily at the Asn-149 position [13].
`
`Oxidation of Cysteine
`
`Under neutral or basic conditions, the free thiol (—SH) group of a cysteine
`is the most reactive moiety of all amino acid components. The disulfide
`(—S—S—) bond formed from the oxidation of two thiol groups results in sig-
`nificant changes in conformation both intramolecularly and intermolecularly.
`Oxidation of the thiol group is promoted at both neutral and basic pH.
`The rate-pH profile for captopril, a quasi dipeptide, showed an increase in
`oxidation rate, starting at pH 5 (Fig. 11). This reaction can be effectively
`retarded by the addition of a metal chelating agent such as EDTA [5].
`A strategy to stabilize proteins that have reactive thiol groups is to re-
`place the cysteine with serine. Serine, a stable amino acid with a hydroxy
`(—OH) group, mimicks the size and polarity of cysteine.
`In many cases,
`these serine mutants retain full biological activities.
`
`Oxidation of Methionine
`
`The methiol (—S—CHg) moiety on methionine is susceptible to oxidation to
`form methylsulfoxide (—SO—CH3) derivatives. Of the three methionines in
`human growth hormone, Met-125 is most reactive, Met-14 is less so and Met-
`170 is not reactive at all [16]. The reactive methionine is likely to be the
`one exposed on a protein surface, and the unreactive one buried within the
`core. Air in the headspace of formulated and freeze-dried growth hormone
`can cause 40% of the growth hormone molecules to be oxidized during a 6 month
`storage period [17].
`
`Disulfide Exchange
`
`Disulfide exchange takes place when a cystine (disulfide) bond is reduced
`to two cysteines; one of them then reacts with another cysteine to form a new
`disulfide (Fig. 12). The reaction is base (hydroxide ion) catalyzed and pro-
`moted by mercaptoethanol, which is sometimes used as an antioxidant. The
`reaction is concentration-dependent, and oligomers are frequently formed as
`a result of disulfide scrambling of bonds between peptide chains.
`Improper
`linkages of disulfide bonds were responsible for a reduction in biological ac-
`tivity of interleukin-2 (IL-2) (Fig. 12). There are three cysteines in IL-2
`at positions 58, 105, and 125, The native protein forms a disulfide linkage
`between the two cysteines at 58 and 105. Cleavage of this disulfide in IL-2
`and the subsequent formation of two less-active isomers with disulfide bonds
`
`293
`
`293
`
`
`
`
`
`
`Wang
`
`294
`
`%deamidation
`
`Years
`
`Figure 10 Deamidation of insulin during storage of Actrapid MC (BP-formula-
`tion) at different temperatures. Each point represents the mean of analyses
`of 4-6 different batches. The desamido insulin content was determined by
`basic disc electrophoresis followed by densitometric scanning of the stained
`gels.
`(From Ref. 11.)
`
`at incorrect positions are promoted by high pH and copper ions [18]. The
`number of disulfide bonds in therapeutic proteins can be shown in the follow-
`ing examples:
`IL-2:1; salmon calcitonin:1; interferon-a:2; human growth
`hormone: 2; insulin: 3; urokinase: 12;
`t-plasminogen activator:17; albumin: 17.
`
`50
`
`nm
`2 30
`
`40
`
`
`7 7w a2
`
`2a=
`
`10
`
`q
`N-C-CH-CH,-SH ——= R-S-S-R
`CH,
`COOH
`
`Figure 11 Oxidation of the thiol moiety in captopril. Reaction rates at 50°C
`is followed by a peak on the HPLC chromatogram. All solutions contain di-
`sodium edetate. Without the chelating agent, reaction rates are much faster.
`(From Ref. 5.)
`
`294
`
`294
`
`
`
`
`
`
`Parenteral Products of Peptides and Proteins
`
`295
`
`Native Interleukin - 2
`
`CoSs
`
`Denoturing
`Conditions
`
`ass =Z \AlkalinepH
`
`ae
`08252 25
`SH
`
`sa)
`
`05)-SH
`ssqes
`
`Less active forms
`
`Figure 12 Schematic representation of the various disulfide-linked isomers of
`interleukin-2 [18]. Under denaturing conditions, i.e., high temperature or
`alkaline pH, a free sulfhydryl group forms a disulfide bond with the sulfhy-
`dryl group from an existing disulfide. The results are proteins with incorrect
`disulfide linkages and "non-native" conformation.
`
`With N-disulfide bonds, there exist (2N)!/2N x N! possible isomers. Cau-
`tion must be exercised to preserve the correct disulfide linkage throughout
`the production process for therapeutic proteins.
`
`Racemization
`
`The racemization reaction is catalyzed by both acid and base. Racemization
`of peptides and proteins results in the formation of diastereomers. As an
`example, racemization of the serine residue in a gonadotropin-releasing hor-
`mone analogue was one of the main degradation reactions when treated with
`0.1 N NaOH at 5°C for 48 hr [19]. Racemization under basic conditions is
`thought to proceed by abstraction of the c-proton from an amino acid in a
`peptide to give a negatively charged planar carbanion. A proton can then be
`returned to this optically inactive intermediate, thus producing a mixture of
`D- and L-enantiomers for the individual amino acid (Fig. 13). Since a pep-
`tide is composed of multiple chiral centers, the product formed is a diastereo-
`mer. Racemization is biologically significant since a peptide comprised of D-
`amino acids is generally metabolized much slower than a naturally occurring
`peptide made of only L-amino acids. For this reason, many new synthetic
`peptides, agonists or antagonists, incorporate D-amino acids. A pH depend-
`ency on racemization was demonstrated in an aqueous degradation study of a
`decapeptide, RS-26306 [10]. At neutral and alkaline pHs, racemization con-
`tributed to more degradation than deamidation (Fig. 14).
`
`295
`
`295
`
`
`
`
`
`
`296
`
`Wang
`
`CHR O
`I
`ll
`
`CHR O
`|
`ll
`
`H
`I
`
`QO
`ll
`
`eioer =_— *NH-C——C- —— SeeS
`ne &
`CHR
`OH
`
`Figure 13 Racemization of an amino acid resulting in a mixture of enantiomers.
`
`Maillard Reaction
`
`In the Maillard reaction, the carbonyl group (RCH=0) from glucose can react
`with the free amino group (e.g., R'NHg in lysine) in a peptide to form a
`Schiff base (RCH=NR'). This reaction is acid catalyzed. During pasteuriza-
`tion (60°C, 10 hr), the loss of antithrombin III biological activity seen in the
`presence of reducing sugars, especially glucose, was attributed to this reac-
`tion [20]. An alternative assessment of poor blood-glucose control in diabetic
`patients is the measurement of glycosylated hemoglobin, which is the product
`of the Maillard reaction. The hospital pharmacist should thus be concerned
`about storing protein in dextrose solution for a prolonged period of time.
`
`oe
`
`io7>
`
`io~*
` Epimerization
`
`pH
`
`Figure 14 Rate-pH profiles for degradation of RS- 26306, N-acetyl-3-(2-naph-
`thyl)-D-Ala- 4- Chloro-D-Phe- 3-(3-pyridinyl)-D-Ala-Ser-Tyr-N, N-diethyl-D-
`Arg-Leu-N, N-diethyl-Arg-Pro-D-Alanylamide, a decapeptide [10].
`
`296
`
`296
`
`
`
`
`
`
`Parenteral Products of Peptides and Proteins
`
`297
`
`Dimerization and Polymerization
`
`Insulin forms a small amount (about 1%) of covalent dimer and polymer during
`2 years storage at 4°C. Production of these species increases dramatically
`with increasing temperature. The dimerization is mainly due to a reaction be-
`tween an N-terminal amino group in one insulin with a carboxamide group of
`a glutamine or an asparagine in another insulin molecule [11].
`The covalent bond reactions aforementioned are examples of key reactions
`that may cause protein or peptide instability during storage. A complex pro-
`tein may undergo many reactions simultaneously.
`
`D. Denaturation
`
`Specific conformation is required for proteins to exert physiological and phar-
`macological activities. Denaturation is a process of altering protein conforma-
`tion. The denaturation process is illustrated in Figure 15 [21]. The folded
`structure (N) unwinds, passes through a critical activated state (A*), to an
`incompletely disordered conformation (ID). Finally, some of the ID forms
`uncoil to become random boils (RC). Heat, extreme pHs, organic solvents,
`high salt concentration, lyophilization, or mechanical stress can denature pro-
`teins.
`
`ID
`
`RC
`Level
`
`Energy
`
`Reaction Coordinate
`
`Figure 15 Schematic illustration of reversible conformational transitions of
`a-lactalbumin [21]. N, Native conformation; A*, critical activated state; ID,
`incompletely disordered conformation; and RC, random coil (fully denatured
`state).
`
`297
`
`297
`
`
`
`
`
`
`298
`
`Wang
`
`Protein conformation refers to the specific tertiary structure, which is
`determined by the primary and secondary structures and the disulfide bonds,
`and is held together by three forces: hydrogen bonding, salt bridges, and
`hydrophobic interactions (Fig. 5).
`Hydrogen bonds are the most important in determining overall protein
`conformation, since they are the major forces that stabilize the secondary
`a-helices and §-sheets, as well as the overall folded structure. Water, the
`ubiquitous medium for most proteins, contributes to this hydrogen bonding.
`Thus, to avoid denaturation in protein during a freeze-drying process, a
`small amount of residual moisture is critical for recovering biological activity
`of the reconstituted protein. Cosolvents such as ethanol and acetone, and
`chaotropic agents such as urea and guanidinium chloride disrupt the hydro-
`gen bonds thus readily denature proteins.
`The ionic side-chains of aspartic acid, glutamic acid, lysine, arginine,
`and histidine, normally found on the surface of the protein, contribute to
`the stability of the native conformation by forming salt bridges. The pH of
`the solvent will determine the charge of the side-chains on these amino acids
`and the extent of ionic bonding. Thus, an extreme pH shift can disrupt these
`salt bridges and lead to denaturation, Further, organic solvents weaken the
`streangth of salt bridges, thus, inappropriate exposure to organic solvent
`can also result in denaturation.
`Because hydrophobic side-chains, i.e., phenyl and hydrocarbon chains,
`are usually tucked inside the protein's globular structure, significant stabil-
`izing effects result from their hydrophobic interactions. This interaction,
`too, is sensitive to the effects of solvents. Disruption of hydrophobic inter-
`actions is often considered the mechanism of denaturation by surfactants,
`heat, mechanical stress, or storage. When a protein molecule unfolds under
`thermal stress, and before it refolds to its natural conformation, hydrophobic
`interaction can result in one of two different types of unnatural conformations.
`In a concentrated protein solution, the hydrophobic groups may interact be-
`tween molecules, resulting in protein aggregation.
`In dilute solutions of pro-
`tein, however, intramolecular interactions are much more likely, and may re-
`sult in a conformation different from the native one. Such unnatural conforma-
`tions may show partial or complete loss of biological activity.
`
`E. Analytical Methods
`
`For protein products a battery of tests are required to assure identity, purity,
`potency, and stability of the formulation. Due to complexity of proteins, as
`of yet no chemical or physical assay can substitute for a bioassay for assess-
`ing the potency. There are in vitro and in vivo bioassays. The in vitro bio-
`assay monitors the response of cells or excised tissue to the stimulation of
`hormones, growth factors, or antibodies. For example, the activity of IL-2
`is determined by measuring [°H]-thymidine incorporated into an IL-2 depend-
`ent murine T-cell line. The in vivo bioassay monitors the pharmacological
`responses of animals to the proteins. For example, the activities of insulin
`and glucagon are determined by measuring post-injection blood sugar in rab-
`bits and cats, respectively. Often, in vitro and in vivo bioassays are not
`sensitive enough to detect subtle changes in protein characteristics. Thus,
`consistency in product performance relies heavily on validated manufacturing
`
`298
`
`298
`
`
`
`
`
`
`Parenteral Products of Peptides and Proteins
`
`299
`
`processes as well as end product testing. When the manufacturing process
`or formulation is significantly changed, the only assured way to prove bio-
`equivalency is to conduct clinical studies in patients.
`The multiplicity of degradation routes for proteins requires combined in-
`formation from various assays to assess the stability of the product. The
`following sections describe common analytical methods employed in protein
`formulations.
`
`UV Spectrometry
`Any protein with at least one tryptophan residue can be detected, at about
`the 0.1 mg/ml level, by ultraviolet (UV) spectroscopy at 270 to 280 nm.
`Phenylalanine and tyrosine contribute about one-sixth the molar absorptivity
`to that of tryptophan. One may use UV for batching assay because of its
`precision, reproducibility, and simplicity. The absorptivity can be deter-
`mined from literature values or derived from the absorbance and mass ob-
`tained from amino-acid composition. For example, the generally accepted value
`for the absorbance at 275 nm of a 1 mg/ml solution of immune globulin is 1.5.
`The UV method can be a rapid assay for in-process quality control. Also,
`because protein aggregates scatter UV light with increased absorbance over
`the range 210 to 350 nm, UV spectrometry can be used to monitor protein ag-
`gregation.
`
`Proteins Assay
`Bradford, Lowry and bicinchoninic acid (BCA) assays are three commonly
`employed colorimetric assays for determination of protein content.
`The Bradford assay is based on a change in the absorption maximum,
`from 465 to 595 nm, of an aromatic sulfonate dye, Coomassie Brilliant Blue
`G-250, in the presence of protein.
`Biuret, with a structure of NHgCONHCONHg, and peptides,