International Journal of Pharmaceutics 203 (2000) 1–60
`
`www.elsevier.com:locate:ijpharm
`
`Review
`Lyophilization and development of solid protein
`pharmaceuticals
`
`Wei Wang*
`
`Biotechnology, Bayer Corporation, 800 Dwight Way, Berkeley, CA 94701, USA
`
`Received 9 September 1999; received in revised form 21 December 1999; accepted 4 April 2000
`
`Abstract
`
`Developing recombinant protein pharmaceuticals has proved to be very challenging because of both the complexity
`of protein production and purification, and the limited physical and chemical stability of proteins. To overcome the
`instability barrier, proteins often have to be made into solid forms to achieve an acceptable shelf life as pharmaceu-
`tical products. The most commonly used method for preparing solid protein pharmaceuticals is lyophilization
`(freeze-drying). Unfortunately, the lyophilization process generates both freezing and drying stresses, which can
`denature proteins to various degrees. Even after successful lyophilization with a protein stabilizer(s), proteins in solid
`state may still have limited long-term storage stability. In the past two decades, numerous studies have been
`conducted in the area of protein lyophilization technology, and instability:stabilization during lyophilization and
`long-term storage. Many critical issues have been identified. To have an up-to-date perspective of the lyophilization
`process and more importantly, its application in formulating solid protein pharmaceuticals, this article reviews the
`recent investigations and achievements in these exciting areas, especially in the past 10 years. Four interrelated topics
`are discussed: lyophilization and its denaturation stresses, cryo- and lyo-protection of proteins by excipients, design
`of a robust lyophilization cycle, and with emphasis,
`instability, stabilization, and formulation of solid protein
`pharmaceuticals. © 2000 Elsevier Science B.V. All rights reserved.
`
`Keywords: Aggregation; Cryoprotection; Denaturation; Excipient; Formulation; Freeze-drying; Glass transition; Stability; Lyopro-
`tection; Residual moisture
`
`1. Introduction
`
`Developing recombinant protein pharmaceuti-
`cals has proved to be very challenging because of
`
`* Tel.: (cid:27)1-510-7054755; fax: (cid:27)1-510-7055629.
`E-mail address: wei.wang.b@bayer.com (W. Wang).
`
`both the complexity of protein production and
`purification, and the limited physical and chemi-
`cal stability of proteins. In fact, protein instability
`is one of the two major reasons why protein
`pharmaceuticals are administered traditionally
`through injection rather than taken orally like
`most small chemical drugs (Wang, 1996). To over-
`
`0378-5173:00:$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.
`PII: S 0 3 7 8 - 5 1 7 3 ( 0 0 ) 0 0 4 2 3 - 3
`
`Amneal v. Cubist
`IPR2020-00193
`Cubist Ex. 2004
`
`

`

`W. Wang :International Journal of Pharmaceutics 203 (2000) 1–60
`
`cal and chemical instabilities and stabilization of
`proteins in aqueous and solid states (Manning et
`al., 1989; Cleland et al., 1993); chemical instability
`mechanisms of proteins in solid state (Lai and
`Topp, 1999); various factors affecting protein sta-
`bility during freeze-thawing, freeze-drying, and
`storage
`of
`solid
`protein
`pharmaceuticals
`(Arakawa et al., 1993); and application of
`lyophilization in protein drug development (Pikal,
`1990a,b; Skrabanja et al., 1994; Carpenter et al.,
`1997; Jennings, 1999). Nevertheless,
`it appears
`that several critical issues in the development of
`solid protein pharmaceuticals have not been fully
`examined,
`including various instability factors,
`stabilization, and formulation of solid protein
`pharmaceuticals.
`the
`To have an up-to-date perspective of
`lyophilization process and more importantly, its
`application in formulating solid protein pharma-
`ceuticals, this article reviews the recent investiga-
`tions and achievements in these exciting areas,
`especially in the past 10 years. Four interrelated
`topics are discussed sequentially,
`lyophilization
`and its denaturation stresses; cryo- and lyo-pro-
`tection of proteins by excipients; design of a ro-
`bust
`lyophilization cycle; and with emphasis,
`instability, stabilization, and formulation of solid
`protein pharmaceuticals.
`
`2. Lyophilization and its denaturation stresses
`
`2.1. Lyophilization process
`
`Lyophilization (freeze-drying) is the most com-
`mon process for making solid protein pharmaceu-
`ticals (Cleland et al., 1993; Fox, 1995). This
`process consists of two major steps: freezing of a
`protein solution, and drying of the frozen solid
`under vacuum. The drying step is further divided
`into two phases: primary and secondary drying.
`The primary drying removes the frozen water and
`the secondary drying removes the non-frozen
`‘bound’ water (Arakawa et al., 1993). The amount
`of non-frozen water for globular proteins is about
`0.3–0.35 g g(cid:28) 1 protein, slightly less than the
`proteins’ hydration shell
`(Rupley and Careri,
`1991; Kuhlman et al., 1997). More detailed analy-
`
`2 c
`
`ome the instability barrier, proteins often have to
`be made into solid forms to achieve an acceptable
`shelf life.
`The most commonly used method for preparing
`solid protein pharmaceuticals is lyophilization
`(freeze-drying). However, this process generates a
`variety of freezing and drying stresses, such as
`solute concentration, formation of ice crystals, pH
`changes, etc. All of these stresses can denature
`proteins to various degrees. Thus, stabilizers are
`often required in a protein formulation to protect
`protein stability both during freezing and drying
`processes.
`Even after successful lyophilization, the long-
`term storage stability of proteins may still be very
`limited, especially at high storage temperatures. In
`several cases, protein stability in solid state has
`been shown to be equal to, or even worse than,
`that in liquid state, depending on the storage
`temperature and formulation composition. For
`example, a major degradation pathway of human
`insulin-like growth factor I (hIGF-I) is oxidation
`of Met59 and the oxidation rate in a freeze-dried
`formulation in air-filled vials is roughly the same
`as that
`in a solution at either 25 or 30°C
`(Fransson et al., 1996). Similarly, the oxidation
`rate of lyophilized interleukin 2 (IL-2) is the same
`as that in a liquid formulation containing 1 mg
`ml(cid:28) 1 IL-2, 0.5% hydroxypropyl-b-cyclodextrin
`(HP-b-CD), and 2% sucrose during storage at 4°C
`(Hora et al., 1992b). At a high water content
`(\50%), the degradation rate of insulin is higher
`in a lyophilized formulation than in a solution
`with similar pH-rate profiles
`in both states
`(Strickley and Anderson, 1996). The glucose-in-
`duced formation of des-Ser
`relaxin in a
`lyophilized formulation is faster than in a solution
`during storage at 40°C (Li et al., 1996). These
`examples indicate that stabilizers are still required
`in lyophilized formulations to increase long-term
`storage stability.
`In the past two decades, numerous studies have
`been conducted in the areas of protein freezing
`and drying, and instability and stabilization of
`proteins during lyophilization and long-term stor-
`age. Many critical issues have been identified in
`this period. These studies and achievements have
`been reviewed elsewhere with emphasis on physi-
`
`

`

`W. Wang :International Journal of Pharmaceutics 203 (2000) 1–60
`
`3
`
`sis of each lyophilization step is provided in Sec-
`tion 4.
`Lyophilization generates a variety of stresses,
`which tend to destabilize or unfold:denature an
`unprotected protein. Different proteins tolerate
`freezing and:or drying stresses to various degrees.
`Freeze-thawing of ovalbumin at neutral pH did
`not cause denaturation (Koseki et al., 1990). Re-
`peated (three times) freeze-thawing of tissue-type
`plasminogen activator (tPA) did not cause any
`decrease in protein activity (Hsu et al., 1995).
`Some proteins can keep their activity both during
`freezing and drying processes, such as a1-an-
`titrypsin in phosphate–citrate buffer (Vemuri et
`al., 1994), porcine pancreatic elastase without ex-
`cipients (Chang et al., 1993), and bovine pancre-
`atic ribonuclease A (RNase A, 13.7 kD) in the
`presence or absence of phosphate (Townsend and
`DeLuca, 1990).
`However, many proteins cannot stand freezing
`and:or drying stresses. Freeze-thawing caused loss
`of activity of
`lactate dehydrogenase
`(LDH)
`(Nema and Avis, 1992; Izutsu et al., 1994b; An-
`dersson and Hatti-Kaul, 1999), 60% loss of L-as-
`paraginase (10 mg ml(cid:28) 1) activity in 50 mM
`sodium phosphate buffer (pH 7.4) (Izutsu et al.,
`1994a),
`and
`aggregation
`of
`recombinant
`hemoglobin (Kerwin et al., 1998). Freeze-drying
`caused 10% loss of the antigen-binding capacity
`of a mouse monoclonal antibody (MN12) (Ress-
`ing et al., 1992), more than 40% loss of bilirubin
`oxidase (BO) activity in the presence of dextran or
`polyvinylalcohol (PVA) (Nakai et al., 1998), loss
`of most b-galactosidase activity at 2 or 20 mg
`ml(cid:28) 1 (Izutsu et al., 1993, 1994a), complete loss of
`phosphofructokinase (PFK) and LDH activity in
`the absence of stabilizers (Carpenter et al., 1986,
`1990; Prestrelski et al., 1993a; Anchordoquy and
`Carpenter, 1996), and dissociation of Erwinia L-
`asparaginase tetramer (135 kD) into four inactive
`subunits (34 kD each) in the absence of any
`protectants (Adams and Ramsay, 1996).
`
`2.2. Denaturation stresses during lyophilization
`
`The lyophilization process generates a variety
`of stresses to denature proteins. These include (1)
`low temperature stress; (2) freezing stresses, in-
`
`in-
`cluding formation of dendritic ice crystals,
`creased ionic strength, changed pH, and phase
`separation; and (3) drying stress (removing of the
`protein hydration shell).
`
`2.2.1. Low temperature stress
`The first quantitative study on low-temperature
`denaturation of a model protein was conducted
`presumably by Shikama and Yamazaki (1961).
`They demonstrated a specific temperature range
`in which ox liver catalase was denatured during
`freeze-thawing. Cold denaturation of catalase at
`8.4 mg ml (cid:28) 1 in 10 mM phosphate buffer (pH 7.0)
`started at (cid:28)6°C. Loss of catalase activity reached
`20% at (cid:28)12°C, remained at this level between
`(cid:28)12°C and near (cid:28)75°C, then decreased gradu-
`ally from (cid:28)75 to (cid:28)120°C. There was almost no
`activity loss between (cid:28)129 and (cid:28)192°C. Similar
`results were also obtained for ovalbumin by
`Koseki et al. (1990). Incubation of frozen ovalbu-
`min solution caused structural change of ovalbu-
`min, as monitored by UV difference spectra,
`which increased with decreasing temperature be-
`tween (cid:28)10 and (cid:28)40°C. Further decrease in
`incubation temperature to (cid:28)80°C caused less
`structural change, and no change at (cid:28)192°C.
`Perlman and Nguyen (1992) reported that inter-
`feron-g(IFN-g) aggregation in a liquid mannitol
`formulation was more severe at (cid:28)20°C than at
`(cid:28)70, 5 and 15°C during storage. To prevent
`freezing-induced complication in studying cold
`protein denaturation, cold and heat denaturation
`of RNase A has been conducted under high pres-
`sure (3 kbar). Under this condition, RNase A
`denatured below (cid:28)22°C and above 40°C (Zhang
`et al., 1995). All these examples are clear indica-
`tion of low temperature denaturation rather than
`a freezing or thawing effect.
`The nature of cold denaturation has not been
`satisfactorily delineated. Since solubility of non-
`polar groups in water increases with decreasing
`temperature due to increased hydration of the
`non-polar groups,
`solvophobic
`interaction in
`proteins weakens with decreasing temperature
`(Dill et al., 1989; Graziano et al., 1997). The
`decreasing solvophobic interaction in proteins can
`reach a point where protein stability reaches zero,
`causing cold denaturation (Jaenicke, 1990). While
`
`

`

`W. Wang :International Journal of Pharmaceutics 203 (2000) 1–60
`
`4 n
`
`ormal or thermal denaturation is entropy-driven,
`cold denaturation is enthalpy-driven (Dill et al.,
`1989; Shortle, 1996). Oligomeric proteins typically
`show cold denaturation, i.e. dissociation of sub-
`unit oligomers, since association is considered to
`be a consequence of hydrophobic interaction
`(Jaenicke, 1990; Wisniewski, 1998). Theoretically,
`the calculated free energy of unfolding (DGunf) for
`proteins has a parabolic relationship with temper-
`ature. This means that a temperature of maximum
`stability exists, and both high and low tempera-
`ture can destabilize a protein (Jaenicke, 1990;
`Kristja´nsson and Kinsella, 1991).
`
`2.2.2. Concentration effect
`Freezing a protein solution rapidly increases the
`concentration of all solutes due to ice formation.
`For example, freezing a 0.9% NaCl solution to its
`eutectic temperature of (cid:28)21°C can cause a 24-
`fold increase in its concentration (Franks, 1990).
`The calculated concentration of small carbohy-
`drates in the maximally freeze-concentrated ma-
`trices (MFCS) is as high as 80% (Roos, 1993).
`Thus, all physical properties related to concentra-
`tion may change, such as ionic strength and rela-
`tive composition of
`solutes due to selective
`crystallization. These changes may potentially
`destabilize a protein.
`Generally, lowering the temperature reduces the
`rate of chemical reactions. However, chemical
`reactions may actually accelerate in a partially
`frozen aqueous solution due to increased solute
`concentration (Pikal, 1999). Due to solute concen-
`tration, the rate of oligomerization of b-glutamic
`acid at (cid:28)20°C was much faster than at 0 or 25°C
`in the presence of a water-soluble carbodiimide,
`1-ethyl-3-(3-dimethylaminopropyl)
`carbodiimide
`(EDAC) (Liu and Orgel, 1997).
`The increase in the rate of a chemical reaction
`in a partially frozen state could reach several
`orders of magnitude relative to that in solution
`(Franks, 1990, 1994).
`The reported oxygen concentration in a par-
`tially frozen solution at (cid:28)3°C is as high as 1150
`times that in solution at 0°C (Wisniewski, 1998).
`The increased oxygen concentration can readily
`oxidize sulphydryl groups in proteins. If a protein
`solution contains any contaminant proteases, con-
`
`centration upon freezing may drastically acceler-
`ate protease-catalyzed protein degradation.
`
`2.2.3. Formation of ice-water interface
`Freezing a protein solution generates an ice-wa-
`ter interface. Proteins can be adsorbed to the
`interface, loosening the native fold of proteins and
`resulting
`in
`surface-induced
`denaturation
`(Strambini and Gabellieri, 1996). Rapid (quench)
`cooling generates a large ice-water interface while
`a smaller interface is induced by slow cooling
`(also see Section 4.2). Chang et al.
`(1996b)
`demonstrated that a single freeze–thaw cycle with
`quench cooling denatured six model proteins, in-
`cluding ciliary neurotrophic factor (CNTF), gluta-
`mate
`dehydrogenase
`(GDH),
`interleukin-1
`receptor antagonist
`(IL-1ra), LDH, PFK, and
`tumor necrosis factor binding protein (TNFbp).
`The denaturation effect of quench cooling was
`greater or equivalent to that after 11 cycles of
`slow cooling, suggesting surface-induced denatu-
`ration. This denaturation mechanism was sup-
`ported by a good correlation (r(cid:30)0.99) found
`between the degree of freeze-induced denaturation
`and that of artificially surface-induced denatura-
`tion. The surface was introduced by shaking the
`protein solution containing hydrophobic Teflon
`beads. In a similar study, a correlation coefficient
`of 0.93 was found between the tendency of freeze
`denaturation and surface-induced denaturation
`for eight model proteins, including aldolase, basic
`fibroblast growth factor (bFGF), GDH, IL-1ra,
`LDH, maleate dehydrogenase (MDH), PFK, and
`TNFbp (Kendrick et al., 1995b). However, there
`was no significant correlation (r(cid:30)0.78) between
`freeze denaturation and thermal denaturation
`temperature (Chang et al., 1996b).
`
`2.2.4. pH changes during freezing
`Many proteins are stable only in a narrow pH
`range, such as low molecular weight urokinase
`(LMW-UK) at pH 6–7 (Vrkljan et al., 1994). At
`extreme pHs, increased electrostatic repulsion be-
`tween like charges in proteins tends to cause
`protein unfolding or denaturation (Goto and
`Fink, 1989; Volkin and Klibanov, 1989; Dill,
`1990). Thus, the rate of protein aggregation is
`strongly affected by pH, such as aggregation of
`
`

`

`W. Wang :International Journal of Pharmaceutics 203 (2000) 1–60
`
`5
`
`interleukin 1b (IL-1b) (Gu et al., 1991), human
`relaxin (Li et al., 1995a), and bovine pancreatic
`RNase A (Townsend and DeLuca, 1990; Tsai et
`al., 1998). Moreover, the solution pH can signifi-
`cantly affect the rate of many chemical degrada-
`tions in proteins (Wang, 1999).
`Freezing a buffered protein solution may selec-
`tively crystallize one buffering species, causing pH
`changes. Na2HPO4 crystallizes more readily than
`NaH2PO4 because the solubility of the disodium
`form is considerably lower than that of
`the
`monosodium form. Because of this, a sodium
`phosphate buffer at pH 7 has a molar [NaH2PO4]:
`[Na2HPO4] ratio of 0.72, but this ratio increases
`to 57 at the ternary eutectic temperature during
`freezing (Franks, 1990, 1993). This can lead to a
`significant pH drop during freezing, which then
`denatures pH-sensitive proteins. For example,
`freezing of a LDH solution caused protein denat-
`uration due to a pH drop from 7.5 to 4.5 upon
`selective crystallization of Na2HPO4 (Anchordo-
`quy and Carpenter, 1996). LDH is a pH-sensitive
`protein and a small drop in pH during freezing
`can partially denature the protein even in the
`presence of stabilizers such as sucrose and tre-
`halose (Nema and Avis, 1992). The pH drop
`during freezing may also explain why freezing
`bovine and human IgG species in a sodium phos-
`phate buffer caused formation of more aggregates
`than in potassium phosphate buffer, because
`potassium phosphate buffer does not show signifi-
`cant pH changes during freezing (Sarciaux et al.,
`1998).
`The pH drop during freezing can potentially
`affect storage stability of
`lyophilized proteins.
`Lyophilized IL-1ra in a formulation containing
`phosphate buffer at pH 6.5 aggregated more
`rapidly than that containing citrate buffer at the
`same pH during storage at 8, 30 and 50°C (Chang
`et al., 1996c). Similarly, the pH drop of a succi-
`nate-containing formulation from 5 to 3–4 during
`freezing appeared to be the cause of less storage
`stability for lyophilized IFN-g than that contain-
`ing glycocholate buffer at the same pH (Lam et
`al., 1996).
`
`2.2.5. Phase separation during freezing
`Freezing polymer solutions may cause phase
`
`separation due to polymers’ altered solubilities at
`low temperatures. Freezing-induced phase separa-
`tion can easily occur in a solution containing two
`incompatible polymers such as dextran and Ficoll
`(Izutsu et al., 1996). During freezing of recombi-
`nant hemoglobin in a phosphate buffer containing
`4% (w:w) PEG 3350, 4% (w:w) dextran T500, and
`150 mM NaCl,
`liquid–liquid phase separation
`occurred and created a large excess of interface,
`denaturing the protein (Heller et al., 1997). Addi-
`tion of 5% sucrose or trehalose could not reverse
`the denaturation effect in the system (Heller et al.,
`1999a).
`Several strategies have been proposed to miti-
`gate or prevent phase separation-induced protein
`denaturation during freezing. These include use of
`alternative salts (Heller et al., 1999a), adjustment
`of the relative composition of polymers to avoid
`or to rapidly pass over a temperature region
`where the system may result
`in liquid–liquid
`phase separation (Heller et al., 1999c), and chemi-
`cal modification of the protein such as pegylation
`(Heller et al., 1999b).
`
`2.2.6. Dehydration stresses
`Proteins in an aqueous solution are fully hy-
`drated. A fully hydrated protein has a monolayer
`of water covering the protein surface, which is
`termed the hydration shell (Rupley and Careri,
`1991). The amount of water in full hydration is
`0.3–0.35 g g (cid:28) 1 protein (Rupley and Careri, 1991;
`Kuhlman et al., 1997). Generally, the water con-
`tent of a lyophilized protein product is less than
`10%. Therefore,
`lyophilization removes part of
`the hydration shell. Removal of the hydration
`shell may disrupt the native state of a protein and
`cause denaturation. A hydrated protein, when
`exposed to a water-poor environment during de-
`hydration, tends to transfer protons to ionized
`carboxyl groups and thus abolishes as many
`charges as possible in the protein (Rupley and
`Careri, 1991). The decreased charge density may
`facilitate protein–protein hydrophobic interac-
`tion, causing protein aggregation.
`Water molecules can also be an integral part of
`an active site(s) in proteins. Removal of these
`functional water molecules during dehydration
`easily inactivates proteins. For example, dehydra-
`
`

`

`W. Wang :International Journal of Pharmaceutics 203 (2000) 1–60
`
`6 t
`
`ion of lysozyme caused loss of activity apparently
`due to removal of those water molecules residing
`functionally in the active site (Nagendra et al.,
`1998).
`Lastly, dehydration during lyophilization may
`cause significant difference in moisture distribu-
`tion in different locations of a product cake. The
`uneven moisture distribution may lead to possible
`localized overdrying, which may exacerbate dehy-
`dration-induced protein denaturation (Pikal and
`Shah, 1997).
`
`2.3. Monitoring protein denaturation upon
`lyophilization
`
`The most common method for monitoring
`protein denaturation upon lyophilization appears
`to be infrared (IR) spectroscopy, although other
`methods have been used such as mass spec-
`troscopy (Bunk, 1997), and Raman spectroscopy
`(Belton and Gil, 1994). In the following section,
`IR methodology is discussed in monitoring
`protein denaturation upon lyophilization followed
`by a discussion on reversibility of protein
`denaturation.
`
`2.3.1. Infrared (IR) spectroscopy
`IR (or FTIR) is probably the most extensively
`used technique today for
`studying structural
`changes in proteins upon lyophilization (Susi and
`Byler, 1986; Dong et al., 1995; Carpenter et al.,
`1998, 1999). The lyophilization-induced structural
`changes can be monitored conveniently in the
`amide I, II, or III region. For lyophilized protein
`samples, residual water up to 10% (w:w) does not
`interfere significantly in the amide I region, a
`frequently used sensitive region for determination
`of secondary structures (Dong et al., 1995). How-
`ever, IR studies on proteins in an aqueous solu-
`tion need either subtraction of water absorption
`or solvent replacement with D2O (Goormaghtigh
`et al., 1994). To make reliable subtraction, high
`protein concentrations (\10 mg ml (cid:28) 1) are rec-
`ommended to increase protein absorption signal,
`and a CaF2 (or BaF2) cell with a path length of 10
`mm or less should be used to control the total
`sample absorbance within 1 (Cooper and Knut-
`son, 1995).
`
`Lyophilization may induce several potential
`changes in the IR spectra of proteins. Disruption
`of hydrogen bonds in proteins during lyophiliza-
`tion generally leads to an increase in frequency
`and a decrease in intensity of hydroxyl stretching
`bands (Carpenter and Crowe, 1989). Unfolding of
`proteins during lyophilization broadens and shifts
`(to higher wave numbers) amide I component
`peaks (Prestrelski et al., 1993b; Allison et al.,
`1996). Lyophilization often leads to an increase in
`b-sheet content with a concomitant decrease in
`a-helix content. Conversion of a-helix to b-sheet
`during lyophilization has been observed in many
`proteins such as tetanus toxoid (TT) in 10 mM
`sodium phosphate buffer (pH 7.3) (Costantino et
`al., 1996), recombinant human albumin (rHA) in
`different buffer
`solutions
`at different pHs
`(Costantino et al., 1995a), hGH at pH 7.8, and
`seven model proteins in water, including bovine
`pancreatic trypsin inhibitor (BPTI), chymotrypsi-
`nogen, horse myoglobin (Mb), horse heart cy-
`tochrome c (Cyt c), rHA, porcine insulin, and
`RNase A (Griebenow and Klibanov, 1995).
`b-sheet
`An
`increase
`in
`content
`during
`lyophilization is often an indication of protein
`aggregation and:or increased intermolecular inter-
`action (Yeo et
`al.,
`1994; Griebenow and
`Klibanov,
`1995; Overcashier
`et
`al.,
`1997).
`Lyophilization-induced increase in b-sheet content
`seems to be a rather general phenomenon as
`lyophilization or air-drying of unordered poly-L-
`lysine induced structural transition to a highly
`ordered b-sheet (Prestrelski et al., 1993b; Wolkers
`et al., 1998b). Such transition has also been ob-
`served in proteins during lyophilization such as
`human insulin in water (pH 7.1) (Pikal and Rigs-
`bee, 1997). The b-sheet structure after lyophiliza-
`tion shows a higher degree of
`intermolecular
`hydrogen bonding because polar groups must sat-
`isfy their H-bonding requirement by intra- or
`intermolecular interaction upon removal of water.
`The intermolecular b-sheet is characterized by two
`major IR bands at about 1617 and 1697 cm(cid:28) 1 in
`solid state, which can be used to monitor protein
`denaturation (Allison et al., 1996). Similarly, the
`relative intensity of a-helix band also can be used
`
`

`

`W. Wang :International Journal of Pharmaceutics 203 (2000) 1–60
`
`7
`
`in this regard (Yang et al., 1999; Heller et al.,
`1999b).
`The extent of changes in overall IR spectrum of
`a protein upon lyophilization reflects the degree of
`protein denaturation. The changes relative to a
`reference spectrum can be measured using a corre-
`lation coefficient (r) as defined by Prestrelski et al.
`(1993a), or the extent of spectral area overlap
`(Heimburg and Marsh, 1993; Allison et al., 1996;
`Kendrick et al., 1996). Using the correlation co-
`efficient, Prestrelski et al. (1993b) were able to
`measure the relative freeze-drying stability of sev-
`eral model proteins, including bFGF, bovine a-
`a-casein,
`IFN-g,
`lactalbumin,
`bovine
`and
`recombinant granulocyte colony-stimulating fac-
`tor (rG-CSF) in the presence of different sugars.
`Nevertheless, Griebenow and Klibanov (1995),
`after analyzing secondary structures of seven
`model proteins upon lyophilization, concluded
`that the correlation coefficient was not highly
`sensitive to structural alterations in proteins. In-
`stead, comparison of overlapping area-normalized
`second-derivative or deconvoluted spectra seemed
`more reliable and objective.
`Recently, IR has been used in real-time moni-
`toring of freezing and dehydration stresses on
`proteins during lyophilization. By this method,
`glucose at 10% was shown to protect lysozyme
`both during the freezing and drying processes
`(Remmele et al., 1997).
`
`2.3.2. Re6ersibility of freezing- or
`lyophilization-induced protein denaturation
`Many proteins denature to various extents
`upon freezing, especially at low concentrations
`(B0.1 mg ml(cid:28) 1). Freezing-induced denaturation
`may or may not be reversible. Freezing lysozyme
`or
`IL-1ra
`caused
`reversible
`denaturation
`(Kendrick et al., 1995a). In contrast, recombinant
`factor XIII (rFXIII, 166 kD) was irreversibly
`denatured upon freezing, and loss of native
`rFXIII at 1 mg ml (cid:28) 1 increased linearly with the
`number of freeze–thaw cycles (Kreilgaard et al.,
`1998b).
`lyophilization-induced denaturation
`Similarly,
`can be either reversible or irreversible. In the
`absence of stabilizers, PFK at 25 mg ml (cid:28) 1 at pH
`7.5 and 8.0 was fully and irreversibly inactivated
`
`upon lyophilization (Carpenter et al., 1993), while
`loss of BO activity in a PVA-containing formula-
`tion was at least partially reversible (Nakai et al.,
`1998). Using IR spectroscopy, Prestrelski et al.
`(1993b) demonstrated that lyophilization-induced
`structural changes were irreversible for bFGF,
`IFN-g, and bovine a-casein, but essentially re-
`versible for G-CSF and bovine a-lactalbumin.
`The extensive aggregation and precipitation of
`IFN-g and casein upon rehydration confirmed the
`irreversibility in structural changes. Therefore,
`lyophilization of proteins may lead to three types
`of behavior, (1) no change in protein conforma-
`tion; (2) reversible denaturation; or (3) irreversible
`denaturation.
`In many cases, IR-monitored structural changes
`during lyophilization seem to be
`reversible.
`Griebenow and Klibanov (1995) demonstrated
`that lyophilization (dehydration) caused signifi-
`cant changes in the secondary structures of seven
`model proteins in the amide III region (1220–
`1330 cm(cid:28) 1), including BPTI, chymotrypsinogen,
`Mb, Cyt c, rHA,
`insulin, and RNase A. The
`structure of almost all proteins became more or-
`dered upon lyophilization with a decrease in the
`unordered structures. Nevertheless, all these struc-
`tural changes were reversible upon reconstitution.
`Other examples of reversible changes in the sec-
`ondary structures of proteins upon lyophilization
`include rHA (Costantino et al., 1995a), Humicola
`lanuginosa lipase (Kreilgaard et al., 1999), IL-2
`(Prestrelski et al., 1995), and lysozyme (Allison et
`al., 1999).
`
`3. Cryo- and lyo-protection of proteins by
`stabilizers
`
`As discussed before, both freezing and dehydra-
`tion can induce protein denaturation. To protect a
`protein from freezing (cryoprotection) and:or de-
`hydration (lyoprotection) denaturation, a protein
`stabilizer(s) may be used. These stabilizers are
`also referenced as chemical additives (Li et al.,
`1995b), co-solutes (Arakawa et al. 1993), co-sol-
`vents
`(Timasheff, 1993, 1998), or
`excipients
`(Wong and Parascrampuria, 1997; Wang, 1999).
`In the following section, a variety of protein
`
`

`

`W. Wang :International Journal of Pharmaceutics 203 (2000) 1–60
`
`8 s
`
`tabilizers are presented for cryo- and lyo-protec-
`tion, followed by discussions of their possible
`stabilization mechanisms.
`
`3.1. Stabilizers for cryo- and lyo-protection
`
`Nature protects life from freezing or osmotic
`shock by accumulating selected compounds to
`high concentrations (\1 M) within organisms.
`These accumulated compounds are known as cry-
`oprotectants and osmolytes, which are preferen-
`tially excluded from surfaces of proteins and act
`as structure stabilizers (Timasheff, 1993). How-
`ever, since the dehydration stress is different from
`those of freezing, many effective cryoprotectants
`or protein stabilizers in solution do not stabilize
`proteins during dehydration (drying). Some even
`destabilize proteins during lyophilization. For ex-
`ample, CaCl2 stabilized elastase (20 mg ml (cid:28) 1) in
`10 mM sodium acetate (pH 5.0), but caused the
`lyophilized protein cake to collapse and lose activ-
`ity (Chang et al., 1993).
`Similarly, effective lyophilization stabilizers (ly-
`oprotectants) may or may not stabilize proteins
`effectively during freezing. Therefore,
`in cases
`when a single stabilizer does not serve as both a
`cryoprotectant and a lyoprotectant, two (or more)
`stabilizers may have to be used to protect proteins
`from denaturation during lyophilization.
`
`3.1.1. Sugars:polyols
`Many sugars or polyols are frequently used
`nonspecific protein stabilizers in solution and dur-
`ing freeze-thawing and freeze-drying. They have
`been used both as effective cryoprotectants and
`remarkable lyoprotectants. In fact, their function
`as lyoprotectants for proteins has long been
`adopted by nature. Anhydrobiotic organisms (wa-
`ter content B1%) commonly contain high con-
`centrations
`(up
`to
`50%)
`of
`disaccharides,
`particularly sucrose or trehalose, to protect them-
`selves (Crowe et al., 1992, 1998).
`The level of stabilization afforded by sugars or
`polyols generally depends on their concentrations.
`A concentration of 0.3 M has been suggested to
`be the minimum to achieve significant stabiliza-
`tion (Arakawa et al., 1993). This has been found
`to be true in many cases during freeze-thawing.
`
`For example, freezing rabbit muscle LDH in wa-
`ter caused 64% loss of protein activity, and in the
`presence of 5, 10 or 34.2% sucrose, the respective
`losses were 27, 12, and 0% (Nema and Avis,
`1992). Other sugars or polyols that can protect
`LDH during freeze-thawing to different degrees
`include lactose, glycerol, xylitol, sorbitol, and
`mannitol, at 0.5–1 M (Carpenter et al., 1990).
`Increasing trehalose concentration gradually in-
`creased the recovery of PFK activity during
`freeze-thawing and the recovery reached a maxi-
`mum of 90% at about 300 mg ml(cid:28) 1 (Carpenter et
`al., 1990). A similar stabilizing trend was also
`observed for sucrose, maltose, glucose, or inositol
`(Carpenter et al., 1986).
`Since freezing is part of the freeze-drying pro-
`cess, high concentrations of sugars or polyols are
`often necessary for lyoprotection. These examples
`include the protection of chymotrypsinogen in the
`presence of 300 mM sucrose (Allison et al., 1996),
`complete inhibition of acidic fibroblast growth
`factor (aFGF) aggregation by 2% sucrose (Volkin
`and Middaugh, 1996), increase in glucose-6-phos-
`phate dehydrogenase (G6PDH) activity from 40
`to about 90% by 5.5% sugar mixture (glu-
`cose:sucrose(cid:30)1:10, w:w) (Sun et al., 1998), com-
`plete recovery of LDH by either 7% sucrose or 7%
`raffinose, a trisaccharide (Moreira et al., 1998),
`significant improvement of PFK recovery by 400
`mM trehalose (Carpenter et al., 1993), and com-
`plete pro