wei wang

International Journal of Pharmaceutics 203 (2000) 1–60

Review

Lyophilization and development of solid proteinpharmaceuticals

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 complexityof protein production and purification, and the limited physical and chemical stability of proteins. To overcome theinstability 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 candenature proteins to various degrees. Even after successful lyophilization with a protein stabilizer(s), proteins in solidstate may still have limited long-term storage stability. In the past two decades, numerous studies have beenconducted in the area of protein lyophilization technology, and instability/stabilization during lyophilization andlong-term storage. Many critical issues have been identified. To have an up-to-date perspective of the lyophilizationprocess and more importantly, its application in formulating solid protein pharmaceuticals, this article reviews therecent investigations and achievements in these exciting areas, especially in the past 10 years. Four interrelated topicsare discussed: lyophilization and its denaturation stresses, cryo- and lyo-protection of proteins by excipients, designof a robust lyophilization cycle, and with emphasis, instability, stabilization, and formulation of solid proteinpharmaceuticals. © 2000 Elsevier Science B.V. All rights reserved.

Keywords: Aggregation; Cryoprotection; Denaturation; Excipient; Formulation; Freeze-drying; Glass transition; Stability; Lyopro-tection; Residual moisture

www.elsevier.com/locate/ijpharm

1. Introduction

Developing recombinant protein pharmaceuti-cals has proved to be very challenging because of

both the complexity of protein production andpurification, and the limited physical and chemi-cal stability of proteins. In fact, protein instabilityis one of the two major reasons why proteinpharmaceuticals are administered traditionallythrough injection rather than taken orally likemost small chemical drugs (Wang, 1996). To over-

* Tel.: +1-510-7054755; fax: +1-510-7055629.E-mail address: [email protected] (W. Wang).

0378-5173/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.

PII: S0 378 -5173 (00 )00423 -3

W. Wang / International Journal of Pharmaceutics 203 (2000) 1–602

come the instability barrier, proteins often have tobe made into solid forms to achieve an acceptableshelf life.

The most commonly used method for preparingsolid protein pharmaceuticals is lyophilization(freeze-drying). However, this process generates avariety of freezing and drying stresses, such assolute concentration, formation of ice crystals, pHchanges, etc. All of these stresses can denatureproteins to various degrees. Thus, stabilizers areoften required in a protein formulation to protectprotein stability both during freezing and dryingprocesses.

Even after successful lyophilization, the long-term storage stability of proteins may still be verylimited, especially at high storage temperatures. Inseveral cases, protein stability in solid state hasbeen shown to be equal to, or even worse than,that in liquid state, depending on the storagetemperature and formulation composition. Forexample, a major degradation pathway of humaninsulin-like growth factor I (hIGF-I) is oxidationof Met59 and the oxidation rate in a freeze-driedformulation in air-filled vials is roughly the sameas that in a solution at either 25 or 30°C(Fransson et al., 1996). Similarly, the oxidationrate of lyophilized interleukin 2 (IL-2) is the sameas that in a liquid formulation containing 1 mgml−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 higherin a lyophilized formulation than in a solutionwith similar pH-rate profiles in both states(Strickley and Anderson, 1996). The glucose-in-duced formation of des-Ser relaxin in alyophilized formulation is faster than in a solutionduring storage at 40°C (Li et al., 1996). Theseexamples indicate that stabilizers are still requiredin lyophilized formulations to increase long-termstorage stability.

In the past two decades, numerous studies havebeen conducted in the areas of protein freezingand drying, and instability and stabilization ofproteins during lyophilization and long-term stor-age. Many critical issues have been identified inthis period. These studies and achievements havebeen reviewed elsewhere with emphasis on physi-

cal and chemical instabilities and stabilization ofproteins in aqueous and solid states (Manning etal., 1989; Cleland et al., 1993); chemical instabilitymechanisms of proteins in solid state (Lai andTopp, 1999); various factors affecting protein sta-bility during freeze-thawing, freeze-drying, andstorage of solid protein pharmaceuticals(Arakawa et al., 1993); and application oflyophilization in protein drug development (Pikal,1990a,b; Skrabanja et al., 1994; Carpenter et al.,1997; Jennings, 1999). Nevertheless, it appearsthat several critical issues in the development ofsolid protein pharmaceuticals have not been fullyexamined, including various instability factors,stabilization, and formulation of solid proteinpharmaceuticals.

To have an up-to-date perspective of thelyophilization process and more importantly, itsapplication 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 interrelatedtopics are discussed sequentially, lyophilizationand 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 solidprotein 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). Thisprocess consists of two major steps: freezing of aprotein solution, and drying of the frozen solidunder vacuum. The drying step is further dividedinto two phases: primary and secondary drying.The primary drying removes the frozen water andthe secondary drying removes the non-frozen‘bound’ water (Arakawa et al., 1993). The amountof non-frozen water for globular proteins is about0.3–0.35 g g−1 protein, slightly less than theproteins’ hydration shell (Rupley and Careri,1991; Kuhlman et al., 1997). More detailed analy-

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 anunprotected protein. Different proteins toleratefreezing and/or drying stresses to various degrees.Freeze-thawing of ovalbumin at neutral pH didnot cause denaturation (Koseki et al., 1990). Re-peated (three times) freeze-thawing of tissue-typeplasminogen activator (tPA) did not cause anydecrease in protein activity (Hsu et al., 1995).Some proteins can keep their activity both duringfreezing and drying processes, such as a1-an-titrypsin in phosphate–citrate buffer (Vemuri etal., 1994), porcine pancreatic elastase without ex-cipients (Chang et al., 1993), and bovine pancre-atic ribonuclease A (RNase A, 13.7 kD) in thepresence or absence of phosphate (Townsend andDeLuca, 1990).

However, many proteins cannot stand freezingand/or drying stresses. Freeze-thawing caused lossof 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−1) activity in 50 mMsodium phosphate buffer (pH 7.4) (Izutsu et al.,1994a), and aggregation of recombinanthemoglobin (Kerwin et al., 1998). Freeze-dryingcaused 10% loss of the antigen-binding capacityof a mouse monoclonal antibody (MN12) (Ress-ing et al., 1992), more than 40% loss of bilirubinoxidase (BO) activity in the presence of dextran orpolyvinylalcohol (PVA) (Nakai et al., 1998), lossof most b-galactosidase activity at 2 or 20 mgml−1 (Izutsu et al., 1993, 1994a), complete loss ofphosphofructokinase (PFK) and LDH activity inthe absence of stabilizers (Carpenter et al., 1986,1990; Prestrelski et al., 1993a; Anchordoquy andCarpenter, 1996), and dissociation of Erwinia L-asparaginase tetramer (135 kD) into four inactivesubunits (34 kD each) in the absence of anyprotectants (Adams and Ramsay, 1996).

2.2. Denaturation stresses during lyophilization

The lyophilization process generates a varietyof stresses to denature proteins. These include (1)low temperature stress; (2) freezing stresses, in-

cluding formation of dendritic ice crystals, in-creased ionic strength, changed pH, and phaseseparation; and (3) drying stress (removing of theprotein hydration shell).

2.2.1. Low temperature stressThe first quantitative study on low-temperature

denaturation of a model protein was conductedpresumably by Shikama and Yamazaki (1961).They demonstrated a specific temperature rangein which ox liver catalase was denatured duringfreeze-thawing. Cold denaturation of catalase at8.4 mg ml−1 in 10 mM phosphate buffer (pH 7.0)started at −6°C. Loss of catalase activity reached20% at −12°C, remained at this level between−12°C and near −75°C, then decreased gradu-ally from −75 to −120°C. There was almost noactivity loss between −129 and −192°C. Similarresults were also obtained for ovalbumin byKoseki 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 −10 and −40°C. Further decrease inincubation temperature to −80°C caused lessstructural change, and no change at −192°C.Perlman and Nguyen (1992) reported that inter-feron-g(IFN-g) aggregation in a liquid mannitolformulation was more severe at −20°C than at−70, 5 and 15°C during storage. To preventfreezing-induced complication in studying coldprotein denaturation, cold and heat denaturationof RNase A has been conducted under high pres-sure (3 kbar). Under this condition, RNase Adenatured below −22°C and above 40°C (Zhanget al., 1995). All these examples are clear indica-tion of low temperature denaturation rather thana freezing or thawing effect.

The nature of cold denaturation has not beensatisfactorily delineated. Since solubility of non-polar groups in water increases with decreasingtemperature due to increased hydration of thenon-polar groups, solvophobic interaction inproteins weakens with decreasing temperature(Dill et al., 1989; Graziano et al., 1997). Thedecreasing solvophobic interaction in proteins canreach a point where protein stability reaches zero,causing cold denaturation (Jaenicke, 1990). While


normal or thermal denaturation is entropy-driven,cold denaturation is enthalpy-driven (Dill et al.,1989; Shortle, 1996). Oligomeric proteins typicallyshow cold denaturation, i.e. dissociation of sub-unit oligomers, since association is considered tobe a consequence of hydrophobic interaction(Jaenicke, 1990; Wisniewski, 1998). Theoretically,the calculated free energy of unfolding (DGunf) forproteins has a parabolic relationship with temper-ature. This means that a temperature of maximumstability exists, and both high and low tempera-ture can destabilize a protein (Jaenicke, 1990;Kristjansson and Kinsella, 1991).

2.2.2. Concentration effectFreezing a protein solution rapidly increases the

concentration of all solutes due to ice formation.For example, freezing a 0.9% NaCl solution to itseutectic temperature of −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 selectivecrystallization. These changes may potentiallydestabilize a protein.

Generally, lowering the temperature reduces therate of chemical reactions. However, chemicalreactions may actually accelerate in a partiallyfrozen aqueous solution due to increased soluteconcentration (Pikal, 1999). Due to solute concen-tration, the rate of oligomerization of b-glutamicacid at −20°C was much faster than at 0 or 25°Cin 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 reactionin a partially frozen state could reach severalorders of magnitude relative to that in solution(Franks, 1990, 1994).

The reported oxygen concentration in a par-tially frozen solution at −3°C is as high as 1150times that in solution at 0°C (Wisniewski, 1998).The increased oxygen concentration can readilyoxidize sulphydryl groups in proteins. If a proteinsolution contains any contaminant proteases, con-

centration upon freezing may drastically acceler-ate protease-catalyzed protein degradation.

2.2.3. Formation of ice-water interfaceFreezing a protein solution generates an ice-wa-

ter interface. Proteins can be adsorbed to theinterface, loosening the native fold of proteins andresulting in surface-induced denaturation(Strambini and Gabellieri, 1996). Rapid (quench)cooling generates a large ice-water interface whilea smaller interface is induced by slow cooling(also see Section 4.2). Chang et al. (1996b)demonstrated that a single freeze–thaw cycle withquench cooling denatured six model proteins, in-cluding ciliary neurotrophic factor (CNTF), gluta-mate dehydrogenase (GDH), interleukin-1receptor antagonist (IL-1ra), LDH, PFK, andtumor necrosis factor binding protein (TNFbp).The denaturation effect of quench cooling wasgreater or equivalent to that after 11 cycles ofslow cooling, suggesting surface-induced denatu-ration. This denaturation mechanism was sup-ported by a good correlation (r=0.99) foundbetween the degree of freeze-induced denaturationand that of artificially surface-induced denatura-tion. The surface was introduced by shaking theprotein solution containing hydrophobic Teflonbeads. In a similar study, a correlation coefficientof 0.93 was found between the tendency of freezedenaturation and surface-induced denaturationfor eight model proteins, including aldolase, basicfibroblast growth factor (bFGF), GDH, IL-1ra,LDH, maleate dehydrogenase (MDH), PFK, andTNFbp (Kendrick et al., 1995b). However, therewas no significant correlation (r=0.78) betweenfreeze denaturation and thermal denaturationtemperature (Chang et al., 1996b).

2.2.4. pH changes during freezingMany 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). Atextreme pHs, increased electrostatic repulsion be-tween like charges in proteins tends to causeprotein unfolding or denaturation (Goto andFink, 1989; Volkin and Klibanov, 1989; Dill,1990). Thus, the rate of protein aggregation isstrongly affected by pH, such as aggregation of


interleukin 1b (IL-1b) (Gu et al., 1991), humanrelaxin (Li et al., 1995a), and bovine pancreaticRNase A (Townsend and DeLuca, 1990; Tsai etal., 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 pHchanges. Na2HPO4 crystallizes more readily thanNaH2PO4 because the solubility of the disodiumform is considerably lower than that of themonosodium form. Because of this, a sodiumphosphate buffer at pH 7 has a molar [NaH2PO4]/[Na2HPO4] ratio of 0.72, but this ratio increasesto 57 at the ternary eutectic temperature duringfreezing (Franks, 1990, 1993). This can lead to asignificant pH drop during freezing, which thendenatures 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 uponselective crystallization of Na2HPO4 (Anchordo-quy and Carpenter, 1996). LDH is a pH-sensitiveprotein and a small drop in pH during freezingcan partially denature the protein even in thepresence of stabilizers such as sucrose and tre-halose (Nema and Avis, 1992). The pH dropduring freezing may also explain why freezingbovine and human IgG species in a sodium phos-phate buffer caused formation of more aggregatesthan in potassium phosphate buffer, becausepotassium phosphate buffer does not show signifi-cant pH changes during freezing (Sarciaux et al.,1998).

The pH drop during freezing can potentiallyaffect storage stability of lyophilized proteins.Lyophilized IL-1ra in a formulation containingphosphate buffer at pH 6.5 aggregated morerapidly than that containing citrate buffer at thesame pH during storage at 8, 30 and 50°C (Changet al., 1996c). Similarly, the pH drop of a succi-nate-containing formulation from 5 to 3–4 duringfreezing appeared to be the cause of less storagestability for lyophilized IFN-g than that contain-ing glycocholate buffer at the same pH (Lam etal., 1996).

2.2.5. Phase separation during freezingFreezing polymer solutions may cause phase

separation due to polymers’ altered solubilities atlow temperatures. Freezing-induced phase separa-tion can easily occur in a solution containing twoincompatible polymers such as dextran and Ficoll(Izutsu et al., 1996). During freezing of recombi-nant hemoglobin in a phosphate buffer containing4% (w/w) PEG 3350, 4% (w/w) dextran T500, and150 mM NaCl, liquid–liquid phase separationoccurred and created a large excess of interface,denaturing the protein (Heller et al., 1997). Addi-tion of 5% sucrose or trehalose could not reversethe denaturation effect in the system (Heller et al.,1999a).

Several strategies have been proposed to miti-gate or prevent phase separation-induced proteindenaturation during freezing. These include use ofalternative salts (Heller et al., 1999a), adjustmentof the relative composition of polymers to avoidor to rapidly pass over a temperature regionwhere the system may result in liquid–liquidphase separation (Heller et al., 1999c), and chemi-cal modification of the protein such as pegylation(Heller et al., 1999b).

2.2.6. Dehydration stressesProteins in an aqueous solution are fully hy-

drated. A fully hydrated protein has a monolayerof water covering the protein surface, which istermed the hydration shell (Rupley and Careri,1991). The amount of water in full hydration is0.3–0.35 g g−1 protein (Rupley and Careri, 1991;Kuhlman et al., 1997). Generally, the water con-tent of a lyophilized protein product is less than10%. Therefore, lyophilization removes part ofthe hydration shell. Removal of the hydrationshell may disrupt the native state of a protein andcause denaturation. A hydrated protein, whenexposed to a water-poor environment during de-hydration, tends to transfer protons to ionizedcarboxyl groups and thus abolishes as manycharges as possible in the protein (Rupley andCareri, 1991). The decreased charge density mayfacilitate protein–protein hydrophobic interac-tion, causing protein aggregation.

Water molecules can also be an integral part ofan active site(s) in proteins. Removal of thesefunctional water molecules during dehydrationeasily inactivates proteins. For example, dehydra-


tion of lysozyme caused loss of activity apparentlydue to removal of those water molecules residingfunctionally in the active site (Nagendra et al.,1998).

Lastly, dehydration during lyophilization maycause significant difference in moisture distribu-tion in different locations of a product cake. Theuneven moisture distribution may lead to possiblelocalized overdrying, which may exacerbate dehy-dration-induced protein denaturation (Pikal andShah, 1997).

2.3. Monitoring protein denaturation uponlyophilization

The most common method for monitoringprotein denaturation upon lyophilization appearsto be infrared (IR) spectroscopy, although othermethods 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 monitoringprotein denaturation upon lyophilization followedby a discussion on reversibility of proteindenaturation.

2.3.1. Infrared (IR) spectroscopyIR (or FTIR) is probably the most extensively

used technique today for studying structuralchanges in proteins upon lyophilization (Susi andByler, 1986; Dong et al., 1995; Carpenter et al.,1998, 1999). The lyophilization-induced structuralchanges can be monitored conveniently in theamide I, II, or III region. For lyophilized proteinsamples, residual water up to 10% (w/w) does notinterfere significantly in the amide I region, afrequently used sensitive region for determinationof secondary structures (Dong et al., 1995). How-ever, IR studies on proteins in an aqueous solu-tion need either subtraction of water absorptionor solvent replacement with D2O (Goormaghtighet al., 1994). To make reliable subtraction, highprotein concentrations (\10 mg ml−1) are rec-ommended to increase protein absorption signal,and a CaF2 (or BaF2) cell with a path length of 10mm or less should be used to control the totalsample absorbance within 1 (Cooper and Knut-son, 1995).

Lyophilization may induce several potentialchanges in the IR spectra of proteins. Disruptionof hydrogen bonds in proteins during lyophiliza-tion generally leads to an increase in frequencyand a decrease in intensity of hydroxyl stretchingbands (Carpenter and Crowe, 1989). Unfolding ofproteins during lyophilization broadens and shifts(to higher wave numbers) amide I componentpeaks (Prestrelski et al., 1993b; Allison et al.,1996). Lyophilization often leads to an increase inb-sheet content with a concomitant decrease ina-helix content. Conversion of a-helix to b-sheetduring lyophilization has been observed in manyproteins such as tetanus toxoid (TT) in 10 mMsodium phosphate buffer (pH 7.3) (Costantino etal., 1996), recombinant human albumin (rHA) indifferent buffer solutions at different pHs(Costantino et al., 1995a), hGH at pH 7.8, andseven model proteins in water, including bovinepancreatic trypsin inhibitor (BPTI), chymotrypsi-nogen, horse myoglobin (Mb), horse heart cy-tochrome c (Cyt c), rHA, porcine insulin, andRNase A (Griebenow and Klibanov, 1995).

An increase in b-sheet content duringlyophilization is often an indication of proteinaggregation and/or increased intermolecular inter-action (Yeo et al., 1994; Griebenow andKlibanov, 1995; Overcashier et al., 1997).Lyophilization-induced increase in b-sheet contentseems to be a rather general phenomenon aslyophilization or air-drying of unordered poly-L-lysine induced structural transition to a highlyordered b-sheet (Prestrelski et al., 1993b; Wolkerset al., 1998b). Such transition has also been ob-served in proteins during lyophilization such ashuman insulin in water (pH 7.1) (Pikal and Rigs-bee, 1997). The b-sheet structure after lyophiliza-tion shows a higher degree of intermolecularhydrogen bonding because polar groups must sat-isfy their H-bonding requirement by intra- orintermolecular interaction upon removal of water.The intermolecular b-sheet is characterized by twomajor IR bands at about 1617 and 1697 cm−1 insolid state, which can be used to monitor proteindenaturation (Allison et al., 1996). Similarly, therelative intensity of a-helix band also can be used


in this regard (Yang et al., 1999; Heller et al.,1999b).

The extent of changes in overall IR spectrum ofa protein upon lyophilization reflects the degree ofprotein denaturation. The changes relative to areference 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 tomeasure the relative freeze-drying stability of sev-eral model proteins, including bFGF, bovine a-lactalbumin, bovine a-casein, IFN-g, andrecombinant granulocyte colony-stimulating fac-tor (rG-CSF) in the presence of different sugars.Nevertheless, Griebenow and Klibanov (1995),after analyzing secondary structures of sevenmodel proteins upon lyophilization, concludedthat the correlation coefficient was not highlysensitive to structural alterations in proteins. In-stead, comparison of overlapping area-normalizedsecond-derivative or deconvoluted spectra seemedmore reliable and objective.

Recently, IR has been used in real-time moni-toring of freezing and dehydration stresses onproteins during lyophilization. By this method,glucose at 10% was shown to protect lysozymeboth during the freezing and drying processes(Remmele et al., 1997).

2.3.2. Re6ersibility of freezing- orlyophilization-induced protein denaturation

Many proteins denature to various extentsupon freezing, especially at low concentrations(B0.1 mg ml−1). Freezing-induced denaturationmay or may not be reversible. Freezing lysozymeor IL-1ra caused reversible denaturation(Kendrick et al., 1995a). In contrast, recombinantfactor XIII (rFXIII, 166 kD) was irreversiblydenatured upon freezing, and loss of nativerFXIII at 1 mg ml−1 increased linearly with thenumber of freeze–thaw cycles (Kreilgaard et al.,1998b).

Similarly, lyophilization-induced denaturationcan be either reversible or irreversible. In theabsence of stabilizers, PFK at 25 mg ml−1 at pH7.5 and 8.0 was fully and irreversibly inactivated

upon lyophilization (Carpenter et al., 1993), whileloss 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-inducedstructural 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 ofIFN-g and casein upon rehydration confirmed theirreversibility in structural changes. Therefore,lyophilization of proteins may lead to three typesof behavior, (1) no change in protein conforma-tion; (2) reversible denaturation; or (3) irreversibledenaturation.

In many cases, IR-monitored structural changesduring lyophilization seem to be reversible.Griebenow and Klibanov (1995) demonstratedthat lyophilization (dehydration) caused signifi-cant changes in the secondary structures of sevenmodel proteins in the amide III region (1220–1330 cm−1), including BPTI, chymotrypsinogen,Mb, Cyt c, rHA, insulin, and RNase A. Thestructure of almost all proteins became more or-dered upon lyophilization with a decrease in theunordered structures. Nevertheless, all these struc-tural changes were reversible upon reconstitution.Other examples of reversible changes in the sec-ondary structures of proteins upon lyophilizationinclude rHA (Costantino et al., 1995a), Humicolalanuginosa lipase (Kreilgaard et al., 1999), IL-2(Prestrelski et al., 1995), and lysozyme (Allison etal., 1999).

3. Cryo- and lyo-protection of proteins bystabilizers

As discussed before, both freezing and dehydra-tion can induce protein denaturation. To protect aprotein from freezing (cryoprotection) and/or de-hydration (lyoprotection) denaturation, a proteinstabilizer(s) may be used. These stabilizers arealso 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


stabilizers are presented for cryo- and lyo-protec-tion, followed by discussions of their possiblestabilization mechanisms.

3.1. Stabilizers for cryo- and lyo-protection

Nature protects life from freezing or osmoticshock by accumulating selected compounds tohigh 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 actas structure stabilizers (Timasheff, 1993). How-ever, since the dehydration stress is different fromthose of freezing, many effective cryoprotectantsor protein stabilizers in solution do not stabilizeproteins during dehydration (drying). Some evendestabilize proteins during lyophilization. For ex-ample, CaCl2 stabilized elastase (20 mg ml−1) in10 mM sodium acetate (pH 5.0), but caused thelyophilized protein cake to collapse and lose activ-ity (Chang et al., 1993).

Similarly, effective lyophilization stabilizers (ly-oprotectants) may or may not stabilize proteinseffectively during freezing. Therefore, in caseswhen a single stabilizer does not serve as both acryoprotectant and a lyoprotectant, two (or more)stabilizers may have to be used to protect proteinsfrom denaturation during lyophilization.

3.1.1. Sugars/polyolsMany sugars or polyols are frequently used

nonspecific protein stabilizers in solution and dur-ing freeze-thawing and freeze-drying. They havebeen used both as effective cryoprotectants andremarkable lyoprotectants. In fact, their functionas lyoprotectants for proteins has long beenadopted 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 orpolyols generally depends on their concentrations.A concentration of 0.3 M has been suggested tobe the minimum to achieve significant stabiliza-tion (Arakawa et al., 1993). This has been foundto 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 thepresence of 5, 10 or 34.2% sucrose, the respectivelosses were 27, 12, and 0% (Nema and Avis,1992). Other sugars or polyols that can protectLDH during freeze-thawing to different degreesinclude lactose, glycerol, xylitol, sorbitol, andmannitol, at 0.5–1 M (Carpenter et al., 1990).Increasing trehalose concentration gradually in-creased the recovery of PFK activity duringfreeze-thawing and the recovery reached a maxi-mum of 90% at about 300 mg ml−1 (Carpenter etal., 1990). A similar stabilizing trend was alsoobserved 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 areoften necessary for lyoprotection. These examplesinclude the protection of chymotrypsinogen in thepresence of 300 mM sucrose (Allison et al., 1996),complete inhibition of acidic fibroblast growthfactor (aFGF) aggregation by 2% sucrose (Volkinand Middaugh, 1996), increase in glucose-6-phos-phate dehydrogenase (G6PDH) activity from 40to about 90% by 5.5% sugar mixture (glu-cose:sucrose=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 400mM trehalose (Carpenter et al., 1993), and com-plete protection of four restriction enzymes by15% trehalose (Colaco et al., 1992). More exam-ples can be found in Table 2.

Lower concentrations of sugars or polyols mayor may not have any significant effect. At 5 to 100mM, neither trehalose nor glucose could protectLDH or PFK to a significant level duringlyophilization (Carpenter et al., 1993). To deter-mine the minimum sugar concentration that offersthe maximum stabilization effect, Tanaka et al.(1991) studied the lyoprotective effect of saccha-rides on the denaturation of catalase duringlyophilization. They demonstrated that saccha-rides protected the protein by direct interactionwith the protein and a concentration of saccha-rides sufficient to form a monomolecular layer onthe protein surface was the minimum to achievethe maximum stabilization. Therefore, the stabi-


lization of catalase was found to depend not onthe bulk concentration of maltose but on theweight ratio of maltose to catalase (Tanaka et al.,1991). Maximum stabilization of catalase was at aratio of about 0.4. A recent study showed thatmaximum protection (about 75% recovery) of L-asparaginase at 1.45 mg ml−1 during lyophiliza-tion was reached at a saccharide concentration ofabout 0.5 mg ml−1, which was about the calcu-lated monosaccharide concentration required tointeract with all exposed highly polar residues ofthe protein (Ward et al., 1999). At this concentra-tion, the weight ratio of saccharide to L-asparagi-nase is 0.34, which is coincidentally very close tothat of maltose to catalase.

On the other hand, increasing sugar/polyol con-centration to a certain level may eventually reacha limit of stabilization or even destabilize aprotein during freeze-drying. For example, actinwas maximally stabilized during lyophilization inthe presence of 5% (w/v) sucrose and a furtherincrease in sucrose concentration to 10% did notimprove the protein stability significantly, whichwas apparently attributable to sticky, pliable, andcollapsed formulation structure (Allison et al.,1998). Increasing trehalose concentration to 150mg ml−1 in a PFK formulation (at 50 mg ml−1,pH 8.0) increased the freeze-drying recovery ofPFK activity to about 65%, but further increasesin trehalose concentration caused a gradual de-crease in recovery of the protein activity (Carpen-ter and Crowe, 1989). At a trehaloseconcentration of 400 mg ml−1, basically no PFKactivity was left after freeze-drying. Since tre-halose at 400 mg ml−1 protected about 90% ofthe protein activity after freeze-thawing, the desta-bilization of PFK at high concentrations of tre-halose occurred in the dehydration step, possiblydue to crystallization of trehalose, preventing req-uisite hydrogen bonding to the dried protein (seeSection 3.2) (Carpenter and Crowe, 1989). A sim-ilar trend was observed in the stabilization ofseveral other proteins during lyophilization in thepresence of increasing concentrations of excipi-ents, including mannitol for L-asparaginase (10 mgml−1) in 50 mM sodium phosphate buffer (pH7.4) (Izutsu et al., 1994b), mannitol for b-galac-tosidase (2 mg ml−1) in 10 mM sodium phosphate

buffer (pH 7.4) (Table 2), mannitol for LDH in 50mM sodium phosphate buffer (pH 7.4) (Izutsu etal., 1994b), and myo-inositol for PFK duringfreeze-thawing and freeze-drying (Table 2). Again,the decreased protein recovery is probably due tocrystallization of these excipients at highconcentrations.

The level of protein protection afforded bydifferent sugars or polyols can be either similar orsignificantly different, depending on the formula-tion composition, concentration and physicalproperties of the stabilizer, and its compatibilitywith the protein. Ward et al. (1999) found thatseveral saccharides, including trehalose, lactose,maltose, sucrose, glucose, and mannitol, displayedsimilar level of protection towards tetrameric L-asparaginase (1.45 mg ml−1) during lyophiliza-tion at saccharide concentrations up to 0.1%. At2%, glucose or lactose protected L-asparaginasefrom dissociation during freeze-drying, but man-nitol did not, possibly due to its crystallizationand loss of intimate interaction with the protein(Adams and Ramsay, 1996). Probably for thesame reason, mannitol at 88 mM inhibited theformation of insoluble hGH aggregates in phos-phate buffer (pH 7.4) at a freezing rate of 50°Cmin−1, but accelerated hGH aggregation at lowerfreezing rates of 0.5 and 5°C min−1 (Eckhardt etal., 1991). In a different study, however, Tanakaet al. (1991) demonstrated that both mannitol andsorbitol could increase the recovery of catalaseactivity during lyophilization to a similar level asthat offorded by maltose. They also showed thatdifferent sugars (maltose, glucose, and mal-totriose) at 1 mg ml−1 could increase the recoveryof catalase activity to the same level (from 35 to90%), but maltopentaose, maltohexaose, and mal-toheptaose were not as effective (Tanaka et al.,1991). The ineffectiveness of larger saccharidessuggests that protein stabilization by sugars maydepend on their glucoside chain lengths, and along chain length may interfere with intermolecu-lar hydrogen-bonding between stabilizing sugarsand proteins.

In many cases, disaccharides appear to be themost effective and universal stabilizers amongsugars and polyols (Arakawa et al., 1993; Carpen-ter et al., 1997). For example, the disaccharides


trehalose, sucrose, maltose, and lactose, were allessentially equivalent to or more effective thanmonosaccharides such as glucose in stabilizingPFK during lyophilization (Crowe et al., 1993b).Trehalose at 400 mM increased the recovery ofPFK activity to greater than 60% duringlyophilization whereas glucose at the same con-centration only recovered less than 5% of theprotein activity (Carpenter et al., 1993). Similarly,the activity of H+-ATPase upon lyophilizationwas increased from 4 to 100, 91, and 84% in thepresence of disaccharides trehalose, maltose, andsucrose, respectively, but only 72 and 37% in thepresence of monosaccharides glucose and galac-tose at 20 mg sugar per mg protein (Sampedro etal., 1998).

Among disaccharides, sucrose and trehalose ap-pear to be the most commonly used. In compari-son to sucrose, trehalose seems to be a preferablelyoprotectant for biomolecules, because it has ahigher glass transition temperature (Crowe et al.,1992, 1996). The higher glass transition tempera-ture of trehalose arises at least partly from theformation of trehalose–protein–water microcrys-tals, preventing water plasticizing the amorphousphase (Librizzi et al., 1999). Other properties oftrehalose are also considered to be advantageous,which include (1) less hygroscopicity, (2) an ab-sence of internal hydrogen bonds, which allowsmore flexible formation of hydrogen bonds withproteins, and (3) very low chemical reactivity(Roser, 1991). To support these arguments, Roser(1991) demonstrated that 35 air-dried restrictionand DNA-modifying enzymes are maximally sta-bilized by 0.3 M trehalose in comparison to othernon-reducing sugars, including sucrose, sorbitol,mannitol, galactitol, etc. as well as reducing sug-ars, including glucose, mannose, galactose, mal-tose, lactose, etc. These advantages of usingtrehalose were later challenged by Levine andSlade (1992), who contended that sucrose couldbe equally effective in protecting biomolecules. Inreality, the relative stabilization effect of these twosugars seems to be depend on both the proteinand sugar concentration. For example, trehaloseat 30 mg ml−1 was more effective in inhibitingIL-6 aggregation during lyophilization than su-crose at the same concentration (Lueckel et al.,

1998b), while sucrose was a better stabilizer thantrehalose during freeze-drying of Humicola lanugi-nosa lipase, a hydrophobic protein (Kreilgaard etal., 1999). Trehalose at 20 mg mg−1 protein wasmore effective than sucrose in stabilizing H+-AT-Pase during lyophilization, but at 510 mg mg−1

protein, sucrose was more effective (Sampedro etal., 1998). In stabilizing G6PDH duringlyophilization, both the glucose/trehalose (1:10,w/w) and glucose/sucrose (1:10, w/w) systemswere shown to be equally effective (Sun andDavidson, 1998).

Not all proteins can be stabilized by sugars/polyols. This is still an unsolved puzzle (Carpenteret al., 1999). For example, sucrose at concentra-tions from 0.1 to 0.5 M showed little effect on theaggregation of recombinant hemoglobin in PBSduring freeze–thaw cycles (Kerwin et al., 1998).Addition of 5% sucrose in MN12 formulation didnot show significant stabilizing effect duringlyophilization (Ressing et al., 1992). Trehalose at5% actually increased the loss of LDH activity inwater from 64 to 74% during freezing (Nema andAvis, 1992). Although the pH of the trehalosesolution decreased during freezing, the pH changeduring freezing could not explain the destabiliza-tion of LDH by trehalose because sucrose, whichstabilized LDH, also caused the same pH change.Therefore, the type of sugar and its subunit orien-tation might have caused the difference in stabiliz-ing LDH (Carpenter et al., 1986).

In a few cases, sugars have to be used withanother excipient(s) to achieve satisfactory proteinstabilization. Carpenter et al. (1986) demonstratedthat freezing rabbit skeletal muscle PFK in liquidnitrogen for 30 s completely inactivated theprotein. Inclusion of 1 mM ZnSO4 or 50 mMsugars (trehalose, sucrose, or maltose) helped toretain less than 13 or 10% of the initial proteinactivity after freeze-thawing, while a combinationof 1 mM ZnSO4 and 50 mM trehalose (sucrose ormaltose) resulted in retention of more than 80%protein activity. More than 85% of protein activ-ity was recovered when ZnSO4 was used withglucose or inositol. Thus, sugars and metal ionshad a synergistic effect in stabilizing PFK duringfreezing. Similarly, neither 10 mM sugar (tre-halose, lactose or mannitol) nor 1% PEG could


improve the lyophilization recovery (46%) ofLDH at 2 mg ml−1 in phosphate buffer (pH 7.5).However, combined use of 1% PEG and 10 mMlactose completely protected the protein from in-activation (Prestrelski et al., 1993a). Also, com-bined use of 1% PEG and sugar (\25 mMtrehalose or glucose) almost completely protectedthe activity of PFK during lyophilization (Car-penter et al., 1993). PEG in these formulationsserved as a lyoprotectant, while the sugars wereused against dehydration denaturation.

3.1.2. PolymersPolymers have been used to stabilize proteins in

solution and during freeze-thawing and freeze-drying (Arakawa et al., 1993). One of the favor-able polymers used in the history of protein drugdevelopment was serum albumin. It has been usedboth as a cryoprotectant and lyoprotectant. Forexample, bovine serum albumin (BSA) at 1%completely protected the activity of rabbit muscleLDH in water during freezing (Nema and Avis,1992). At much lower concentrations between0.05 and 0.1% (w/v), human serum albumin(HSA), due to its effective inhibition of proteinsurface adsorption and general stabilization ofproteins during lyophilization, was used in formu-lating freeze-dried hydrophobic cytokines, such asinterleukin-1a (IL-1a), IL-1b, IL-3, andmacrophage colony stimulating factor (MCSF)(Dawson, 1992). Increasing BSA concentrationsto 0.05% gradually increased the activity recoveryof LDH at 25 mg ml−1 from about 30 to 100%during freeze-thawing and to about 80% duringfreeze-drying (Anchordoquy and Carpenter,1996). Many protein products on the market,such as Betaseron®, Epogen®, Kogenate®, andRecombinate™ contain albumin (Physicians’Desk Reference, 1999). However, the ever-increas-ing concern about the potential contamination ofserum albumin with blood-borne pathogens limitsits future application in protein products. There-fore, rHA has been recommended recently toreplace serum albumin as a protein stabilizer(Tarelli et al., 1998). Nevertheless, the ultimatesolution is to develop albumin-free formulationsfor protein pharmaceuticals.

In addition to albumin, other polymers alsohave been used. The level of protein stabilizationafforded by these polymers depends on structureand concentration of both the polymer and theprotein. For example, dextran (5%), PVA (2.5%),hydroxypropyl methylcellulose (HPMC) (1%), orgelatin (0.5%) reduced the loss of rabbit muscleLDH activity in water during freezing from 64 to24%, 24, 18, and 9%, respectively (Nema andAvis, 1992). LDH activity was also protectedduring lyophilization in the presence of differentconcentrations of polyethyleneimine (Anderssonand Hatti-Kaul, 1999). While ovalbumin at 0.01%had little effect on the stability of catalase (8.4 mgml−1 in 10 mM phosphate buffer, pH 7.0) duringfreezing, gelatin at the same concentration com-pletely protected the protein activity (Shikamaand Yamazaki, 1961). Polyvinylpyrrolidone (PVP)(40 kD) increased both the freeze-thawing andfreeze-drying recovery of LDH in a concentra-tion-dependent manner (Anchordoquy and Car-penter, 1996). Addition of 2% dextran (192 kD)into a sucrose-containing actin formulation sig-nificantly increased the protein stability duringlyophilization (Allison et al., 1998). Hydroxyethylcellulose (HEC) at 1% completely inhibitedlyophilization-induced aggregation of aFGF at100 mg ml−1 in PBS containing 33 mg ml−1

heparin, although reconstitution time was in-creased significantly (Volkin and Middaugh,1996).

Stabilization of proteins by polymers can gener-ally be attributed to one or more of these polymerproperties: preferential exclusion, surface activity,steric hindrance of protein–protein interactions,and/or increased solution viscosity limitingprotein structural movement. In recent years, ad-ditional properties of polymers have been impli-cated in stabilizing proteins during freeze-thawingand freeze-drying. Polymers such as dextran havebeen reported to stabilize proteins by raising theglass transition temperature of a protein formula-tion significantly and by inhibiting crystallizationof small stabilizing excipients such as sucrose(Skrabanja et al., 1994). PEG 3350 or dextranT500 at 4% (w/w) has been found to inhibit a pHdrop during freezing of a phosphate-buffered so-lution by inhibiting crystallization of disodium


phosphate (Heller et al., 1997). Probably by thesame mechanism, BSA or PVP (40 kD) at 10%dramatically inhibited the pH drop during freez-ing of a buffered LDH solution (Anchordoquyand Carpenter, 1996). At least partly due to thisinhibition effect, both BSA and PVP increased thefreeze–thaw recovery of the protein in a concen-tration-dependent manner. The inhibition of crys-tallization of small molecules is apparently due topolymer-induced viscosity increase (Slade et al.,1989).

On the other hand, polymers may cause phaseseparation during freezing, adversely affectingprotein stability (see Section 2.2). Certain poly-mers may destabilize proteins during lyophiliza-tion due to steric hindrance, preventing efficienthydrogen bonding with proteins. Dextran (40 kD)at concentrations of up to 100 mg ml−1 failed toinhibit dehydration-induced unfolding oflysozyme because of its inability to form adequatehydrogen bonding with the protein (Allison et al.,1999). Similarly, this compound could hardly pre-vent formation of b-sheets in poly-L-lysine duringdehydration (Wolkers et al., 1998b). In fact, Dex-tran (162 kD) at 5% (w/v) was shown to destabi-lize Humicola lanuginosa lipase duringlyophilization, as determined by IR (Kreilgaard etal., 1999).

3.1.3. Protein itselfProtein aggregation in solution is generally con-

centration-dependent. It has been suggested thatincreasing protein concentration to higher than0.02 mg ml−1 may facilitate potential proteinaggregation (Ruddon and Bedows, 1997). Increas-ing protein concentration increases aggregation ofmany proteins in solution, such as LMW-UK inthe range 0.2–0.9 mg ml−1 (Vrkljan et al., 1994),IL-1b in the range 100–500 mg ml−1 (Gu et al.,1991), apomyoglobin in the range 4–12 mg ml−1

in the presence of 2.4 M urea (De Young et al.,1993), and insulin (Brange et al., 1992a).

In contrast, proteins at higher concentrationsare often more resistant against both freezing-and lyophilization-induced protein denaturation/aggregation. The activity recovery of many labileproteins after freeze-thawing correlates directlywith initial protein concentration (Allison et al.,

1996). For example, increasing initial concentra-tion of rhFXIII from 1 to 10 mg ml−1 increasedthe recovery of native rhFXIII during repeatedfreeze-thawing (Kreilgaard et al., 1998b). The re-covery of LDH activity gradually increased from6% at a protein concentration of 10 mg ml−1 toabout 65% at concentrations above 175 mg ml−1

after freeze-thawing (Carpenter et al., 1990). Upto about 90% LDH activity was recovered whenthe concentration was increased to 500 mg ml−1

(Anchordoquy and Carpenter, 1996). Koseki et al.(1990) demonstrated that increasing the ovalbu-min concentration in the range 0.5–2.5 mg ml−1

at pH 1.9 decreased the structural changes of thefreeze-treated (−40°C) protein, as measured byUV.

Similarly, the lyophilization recovery of PFKactivity at 25 and 40 mg ml−1 was 34 and 64%,respectively, in the presence of 200 mM trehalose(Carpenter et al., 1987). Increasing the concentra-tion of rabbit muscle LDH from 10 to 500 mgml−1 gradually increased the activity recoveryfrom less than 20% to about 60% duringlyophilization (Anchordoquy and Carpenter,1996). Increasing the concentration of bovine andhuman IgG species markedly decreasedlyophilization-induced protein aggregation (Sar-ciaux et al., 1998). Certain proteins, however, donot show this concentration-dependent protec-tion. The percentage of lyophilization-induced de-naturation of catalase in the absence of astabilizer was determined to be about 65%, inde-pendent of the protein concentration in the range1–5000 mg ml−1 (Tanaka et al., 1991).

The mechanisms of proteins’ self-stabilizationduring freezing and/or lyophilization have notbeen clearly delineated. Proteins are polymers,and therefore, at least some of the above-dis-cussed stabilization mechanisms for polymers maybe applicable to proteins’ self-stabilization. Re-cently, two hypotheses have been reiterated toexplain the concentration-dependent protein sta-bilization upon freezing (Allison et al., 1996).First, unfolding of proteins at high concentrationsduring freezing may be temporarily inhibited bysteric repulsion of neighboring protein molecules.Second, the surface area of ice-water interfaceformed upon freezing is finite, which limits the


amount of protein to be accumulated and dena-tured at the interface. In addition, favorableprotein–protein interactions (possible formationof dimers or multimers) may contribute to theincreased protein stability at high concentrations,as observed for thermophilic proteins (Mozhaevand Martinek, 1984).

3.1.4. Non-aqueous sol6entsNon-aqueous solvents generally destabilize

proteins in solution. At low concentrations certainnon-aqueous solvents may have a stabilizing ef-fect. These stabilizing non-aqueous solvents in-clude polyhydric alcohols such as PEGs, ethyleneglycol, and glycerol and some polar and aproticsolvents such as dimethylsulphoxide (DMSO) anddimethylformamide (DMF) (Volkin andKlibanov, 1989; Carpenter et al., 1991).

In fact, polyhydric alcohols are among the com-monly used and effective cryoprotectants. Forexample, in the presence of 0.2 M PEG 400, theloss of rabbit muscle LDH activity upon freezingwas reduced from 64 to 15% (Nema and Avis,1992). LDH can also be protected from freeze-thawing denaturation to different degrees by eth-ylene glycol or 2-methyl-2,4-pentanediol(Carpenter et al., 1990). PEG at 1–10% (w/v)completely protected both LDH and PFK at 25mg ml−1 (at pH 7.5 and 8.0, respectively) duringfreeze-thawing, although they were not effectivestabilizers during freeze-drying (Carpenter et al.,1993). Glycerol at 0.3% (v/v) prevented freezingdenaturation of ovalbumin (0.1%) (Koseki et al.,1990) and at 1 M, increased the recovery ofcatalase activity upon freezing from 80 to 95%(Shikama and Yamazaki, 1961).

Cryoprotection of proteins by these non-aqueous solvents may be pH-dependent. Ethyleneglycol stabilized RNase A at pH 2.3 but destabi-lized it at pH 5.5 (Arakawa et al., 1991). This ispartly because proteins may tolerate freezing de-naturation to different degrees at different pHs.

3.1.5. SurfactantsThe formation of ice-water interfaces during

freezing may cause surface denaturation ofproteins (see Section 2.2). Surfactants may dropsurface tension of protein solutions and reduce

the driving force of protein adsorption and/oraggregation at these interfaces. Low concentra-tions of nonionic surfactants are often sufficientto serve this purpose due to their relatively lowcritical micelle concentrations (CMC) (Bam et al.,1995). Other stabilization mechanisms were alsoproposed, such as assistance in protein refoldingduring thawing and protein binding, which mayinhibit protein–protein interactions (Carpenter etal., 1999).

Tween 80 is one of the commonly used surfac-tants for protein stabilization during freezing.Tween 80 at concentrations of ]0.01% protectedboth LDH and GDH from denaturation duringquench freezing and thawing (Chang et al.,1996b). The freeze–thaw recovery of LDH activ-ity was increased from 36 to 57 and 65% in thepresence of 0.002 and 0.005% Tween 80, respec-tively (Nema and Avis, 1992). Maximum freeze–thaw recovery (about 80%) of LDH activity wasreached in the presence of 0.05% Tween 80.Tween 80 at concentrations from 0.005 to 0.01%also protected several other proteins from freezingdenaturation, including TNFbp, IL-1ra, bFGF,MDH, aldolase, and PFK (Kendrick et al.,1995b).

Other nonionic and ionic surfactants have alsobeen reported in cryoprotection of proteins. Thefollowing surfactants protected LDH from freez-ing denaturation to various degrees, Brij 35; Brij30 (polyoxyethylene lauryl ether); Lubrol-px; Tri-ton X-10; Pluronic F127 (polyoxyethylene-poly-oxypropylene copolymer); and SDS (Nema andAvis, 1992; Chang et al., 1996b). SDS at 0.5 mMincreased the activity recovery of catalase (8.4 mgml−1 in 10 mM phosphate buffer, pH 7.0) from80 to 90% upon freezing (Shikama and Yamazaki,1961).

3.1.6. Amino acidsCertain amino acids can be used as cryoprotec-

tants and/or lyoprotectants. For example, freezingrabbit skeletal muscle PFK in liquid nitrogen for30 s inactivated the protein completely, and sev-eral amino acids, including glycine, proline, or 4hydroxyproline, significantly increased the recov-ery of the protein activity (Carpenter et al., 1986).Glycine at low concentrations has been shown to


suppress the pH change in 10 or 100 mM sodiumphosphate buffer during freezing (Pikal and Car-penter, 1998). Therefore, amino acids may protectproteins from freezing denaturation at least partlyby reducing the rate and extent of buffer saltcrystallization.

As lyoprotectants, several amino acids in-creased the lyophilization recovery of LDH from22 to about 39–100%, including proline, L-serine,sodium glutamate, alanine, glycine, lysine hydro-chloride, sarcosine, g-aminobutyric acid (Carpen-ter et al., 1990). Glycine alone or in combinationwith mannitol inhibited aggregation of an anti-body-vinca conjugate during lyophilization (Royet al., 1992). LDH activity was increased by 20%during vacuum-drying in the presence of pheny-lalanine:arginine:H3PO4 (1:1:0.5 molar ratio)(Mattern et al., 1999).

3.1.7. Miscellaneous excipientsSalts and amines have been used as cryoprotec-

tants. LDH activity can be protected to variousdegrees upon freezing in the presence of potas-sium phosphate, sodium acetate, ammonium sul-fate, magnesium sulfate, sodium sulfate,trimethylamine N-oxide, or betaine (Carpenter etal., 1990). Increasing the potassium phosphateconcentration from 10 mM to 1 M increased therecovery of LDH upon freezing from less than20% to more than 80% (Arakawa et al., 1993).

Metal ions can protect certain proteins duringlyophilization. In the presence of 100 mM sugarssuch as trehalose, maltose, sucrose, glucose orgalactose, some divalent metal ions improved therecovery of PFK activity (at 40 mg ml−1 in 1 mMsodium borate, pH 7.8) during lyophilization in aconcentration-dependent manner. The relative ef-fectiveness of these metal ions was apparently inthe following order: Zn2+\Cu2+\Ca2+,Mn2+\Mg2+ (Carpenter et al., 1987).

The activity of LDH can be protected to differ-ent degrees during lyophilization in the presenceof some amphiphilic excipients, including HP-b-CD, 3-[(3-cholamidepropyl)-dimethylammonio]-1-propanesulfate (CHAPS), sodium cholate, sucrosemonolaurate (Izutsu et al., 1995). Combinationsof sucrose and these amphiphilic excipients in-creased the protein stability synergistically.

Recently, Ramos et al. (1997) demonstratedthat 2-O-b-mannosylglycerate at 500 mM in-creased the freeze-drying recovery of LDH activ-ity (at 50 ug ml−1) from 12 to 85% whiletrehalose only increased the recovery to 54%.

3.2. Mechanisms of protein stabilization duringlyophilization

Since freezing and drying stresses imposed onproteins during lyophilization are different, mech-anisms of protein stabilization by excipients arenot the same in the two stages of lyophilization.

3.2.1. Mechanisms of cryoprotectionOne of the most widely accepted protein stabi-

lization mechanisms in liquid state is preferentialinteraction. Preferential interaction means that aprotein prefers to interact with either water or anexcipient(s) in an aqueous solution. In the pres-ence of a stabilizing excipient, the protein prefersto interact with water (preferential hydration) andthe excipient is preferentially excluded from thedomain of the protein (preferential exclusion). Inthis case, proportionally more water moleculesand fewer excipient molecules are found at thesurface of the protein than in the bulk. Therefore,preferential exclusion of an excipient is usuallyassociated with an increase in the surface tensionof water. Detailed discussion of this stabilizationmechanism can be found elsewhere (Arakawa etal., 1991, 1993; Timasheff, 1993; Lin andTimasheff, 1996; Timasheff, 1998).

The preferential interaction mechanism appliesequally well during freeze–thaw processes (Car-penter et al., 1991; Arakawa et al., 1993; Crowe etal., 1993b). Protein stabilizers, which are excludedfrom protein surface in solution, can also stabilizeproteins during freezing. Nema and Avis (1992)examined the stabilizing effect of 13 cryoprotec-tants on the recovery of rabbit muscle LDH activ-ity, including trehalose (5%), mannitol (5%),sucrose (5, 10, 34.2%), Brij 30 (polyoxyethylenelauryl ether, 0.05%), Tween 80 (0.002–1%),Pluronic F127 (1%), HPMC (1%), PVP (2.5%),PEG 400 (0.2 M), gelatin (0.5%), BSA (1%),b-cyclodextrin (0.9%) and dextran (5%). Theyfound that the cryoptotectants that increased the


stability of LDH in solution at room temperaturealso improved the recovery of protein activityafter freeze–thaw. However, no apparent correla-tion was found between the increase in surfacetension induced by the cryoprotectants and theirprotective effect on protein recovery duringfreeze–thaw cycles. Based on this study, severalother stabilization mechanisms were postulated,including modification of the size of ice crystals,reduction (instead of elevation) of surface tension,and restriction of diffusion of reacting molecules.Supposedly, reduction of surface tension is howmost surfactants stabilize proteins during freezing.

Besides polymers, many cryoprotectants can in-crease the viscosity of a solution, restricting diffu-sion of reacting molecules. In fact, the differencein solution viscosity has explained why trehaloseis apparently more effective than sucrose, maltose,glucose, or fructose in stabilizing liquid pyrophos-phatase and G6DPH (Sola-Penna and Meyer-Fer-nandes, 1998). On top of this, concentration of allsolutes during freezing increases the solution vis-cosity rapidly. Therefore, the rate of a chemicalreaction may increase initially due to concentra-tion of all solutes but then drops gradually as theviscosity increases (Pikal, 1999). The rate of achemical reaction is minimized at the glassy statewhen the viscosity is increased to 1012 Pa·s (An-gell, 1995). In addition to viscosity increase, someof these cryoprotectants stabilize proteins by sup-pressing pH changes during freezing (Anchordo-quy and Carpenter, 1996).

The preferential interaction mechanism doesnot fully explain protein cryoprotection by poly-mers or by proteins themselves at high concentra-tions. These different mechanisms have beenaddressed in Section 3.1.

3.2.2. Mechanisms of lyoprotectionDuring lyophilization, the preferential interac-

tion mechanism is no longer applicable becausethe hydration shell of proteins is removed (Car-penter et al., 1993; Crowe et al., 1993b; Allison etal., 1996). Thus, many excipients that stabilizeproteins in solution do not offer the same effectduring lyophilization. For example, KCl at 500mM effectively protected LDH from thermal in-activation at 50°C, but did not protect the protein

activity at all during lyophilization (Ramos et al.,1997).

One major mechanism of protein stabilizationby lyoprotectants is the formation of an amor-phous glass during lyophilization (Roser, 1991;Franks, 1994; Fox, 1995). Formation of a glassincreases the viscosity to 1012 Pa s (1013 P)(Angell, 1995). It is the extreme viscosity at theglassy state, that increases protein stability byslowing down interconversion of conformationalsubstates and conformational relaxation of aprotein (Hagen et al., 1995, 1996). This stabiliza-tion mechanism explains the retention of G6PDHactivity during freeze-drying (Sun et al., 1998).Amorphous materials are structurally more simi-lar to a liquid than crystalline materials (Taylorand Zografi, 1998b). Freeze-dried amorphous in-sulin is far more stable than crystalline insulinagainst deamidation and dimer formation at dif-ferent water contents up to 15% (Pikal and Rigs-bee, 1997). Izutsu et al. (1995) studied the effect ofamphiphilic excipients on freeze-drying of LDHand found that only those that remain amorphousin the solid state protected the enzyme duringfreeze-drying. These excipients, including HP-b-CD, CHAPS, sodium cholate, sucrose monolau-rate, showed a concentration-dependentstabilization effect during freeze-drying.

A glass can be roughly divided into two types:fragile and strong. The viscosity of a fragile glassincreases more deeply than a stronger glass for agiven temperature drop below the glass transitiontemperature (Angell, 1995). Therefore, excipientsforming fragile glasses are better stabilizing agents(Hatley, 1997). Both sucrose and trehalose canform a fragile glass (Hatley, 1997; Duddu et al.,1997).

Another interrelated stabilization mechanism isthe water replacement hypothesis (Crowe et al.,1993a; Allison et al., 1996, 1998). This mechanisminvolves the formation of hydrogen bonds be-tween a protein and an excipient(s) at the end ofthe drying process to satisfy the hydrogen bond-ing requirement of polar groups on the proteinsurface (Carpenter and Crowe, 1989; Carpenter etal., 1990). These excipients preserve the nativestructures of proteins by serving as water substi-tutes (Carpenter et al., 1990; Arakawa et al., 1991;


Carpenter et al., 1993; Prestrelski et al., 1995). Inthis way, intra- or interprotein hydrogen bondingmay be prevented during dehydration (Leslie et al.,1995; Cardona et al., 1997). Therefore, stabilizationof proteins requires hydrogen bonding with anexcipient(s) during freeze-drying or dehydration(Carpenter and Crowe, 1989; Arakawa et al., 1991;Carpenter et al., 1991; Crowe et al., 1998).

Since an amorphous state of proteins and stabi-lizers allows maximal H-bonding between proteinand stabilizer molecules, crystallization of anyamorphous protein stabilizers during lyophiliza-tion often causes protein destabilization due toinefficient hydrogen bonding. Mannitol can easilybe crystallized and its crystallization is apparentlyresponsible for the destabilization of some proteinsduring lyophilization. The aggregation of IL-6during lyophilization could not be inhibited effec-tively in a formulation containing only mannitol(Lueckel et al., 1998b). In the presence of 1% PEG,increasing the mannitol concentration above 10mM reduced the activity of LDH and PFK, possi-bly due to crystallization of mannitol (Carpenter etal., 1993). Mannitol at 300 mM destabilized Humi-cola lanuginosa lipase during lyophilization andDSC analysis indicated that 85% mannitol wascrystallized during lyophilization (Kreilgaard et al.,1999).

Although it was debatable whether or not hydro-gen bond was indeed formed between trehalose andlysozyme upon lyophilization (Belton and Gil,1994), many studies have confirmed hydrogenbonding by IR spectroscopy between carbohy-drates and freeze-dried proteins, such as lysozyme,BSA, and PFK (Carpenter and Crowe, 1989;Crowe et al., 1993b; Remmele et al., 1997; Allisonet al. 1999) and bFGF, g-IFN, recombinant G-CSF, bovine a-lactalbumin, and bovine a-casein(Prestrelski et al., 1993b). The degree of structuralprotection of lysozyme by sucrose and trehalosespectra was shown to correlate with the extent ofhydrogen bonding between the sugars and theprotein (Allison et al., 1999). Hydrogen bondinghas also been demonstrated between sucrose andother non-protein polymers, such as poly-L-lysine(Wolkers et al., 1998b), and PVP (Taylor andZografi, 1998b). Different excipients may formhydrogen bonds with proteins to different extents

because of their structural differences. Sucrose hasbeen found to form hydrogen bonds with lysozymeto a greater extent than trehalose (Allison et al.,1999) and with PVP than both trehalose andraffinose (Taylor and Zografi, 1998b). The differ-ence among sugars in stabilization of proteins maybe partially due to the difference in the extent andintimacy of hydrogen bond formation.

In addition to glass formation, many excipients,especially polymers, can stabilize proteins by in-creasing Tg of protein formulations, since higherTgs generally result in more stable protein formula-tions during lyophilization. For example,Costantino et al. (1998b) examined six stabilizers(lactose, trehalose, cellobiose, mannitol, sorbitol,and methyl a-D-mannopyranoside) duringlyophilization of rhGH and found that the higherthe Tg of the stabilized formulation, the greater thedegree of structural (such as a-helix) preservationin the co-lyophilizate with less protein aggregation.In general, larger carbohydrates form a glass morereadily with a higher Tg than smaller ones, but havemore steric hindrance interfering with intimatehydrogen bonding with a dried protein (Crowe etal., 1993b). Therefore, selection of such an excipientneeds balancing both the formation of a glass witha high Tg and intimacy of hydrogen bonding.

Other mechanisms of protein stabilization alsoseem operable. Sugars may stabilize proteins byinhibiting crystallization of other excipients such asPEGs during lyophilization (Izutsu et al., 1995), byinhibiting acute lyophilization-induced protein un-folding such as rhIL-1ra (Chang et al., 1996a), orby preserving a protein’s internal mobility such assperm whale Mb (Sastry and Agmon, 1997).Polyelectrolytes can stabilize a protein duringlyophilization by forming multiple electrostaticinteractions with the protein (Gibson, 1996).

4. Design of a robust lyophilization cycle — astep-by-step analysis

The purpose of designing a robust lyophiliza-tion cycle for protein pharmaceuticals is to obtaina consistent, stable, and esthetically acceptableproduct. To achieve this goal, a number of


parameters that directly determine or characterizea lyophilization cycle need to be determined ordefined. These parameters should include glasstransition temperature (Tg% )/collapse temperature(Tcol), cooling rate, drying rate, and residual mois-ture content.

4.1. Characterization of protein formulations priorto lyophilization

In addition to glass transition temperature (Tg% )/collapse temperature (Tcol), several other criticaltemperatures, including crystallization tempera-ture (Tcry), eutectic temperature (Teut), and devit-rification temperature (Tdev), should bedetermined in order to design a robust lyophiliza-tion cycle. These temperatures are mostly deter-mined by thermal analysis such as DSC, electricalresistance measurements, and direct microscopicobservation.

4.1.1. Glass transition temperature (Tg% ) andcollapse temperature (Tcol)

Ice formation during cooling of a protein solu-tion concentrates all solutes. Solute concentration

eventually changes the solution from a viscousliquid to a brittle glass, which contains about20–50% water (Pikal, 1990b; Hatley et al., 1996).The temperature of this reversible transition forthe maximally freeze-concentrated solution istermed glass transition temperature, Tg% . This tem-perature is also called the temperature of vitreoustransformation (Rey, 1999). Tg% is used to differen-tiate this transition from the softening point of atrue glass transition, Tg of a pure polymer. Tg% isone of the most important parameters for opti-mization of a lyophilization process (Franks,1990).

The collapse temperature (Tcol) is the tempera-ture at which the interstitial water in the frozenmatrix becomes significantly mobile (Jennings,1999). Tcol is closely related to Tg% . In fact, Tcol hasbeen considered to be equivalent to Tg of anamorphous system or to the eutectic melting tem-perature of a crystalline system (Slade et al., 1989;Pikal, 1990a,b). Recent literature indicates thatthe Tcol of many small carbohydrates is consis-tently higher than their Tg% by about 12 K (Sun,1997). The discrepancy between Tg% and Tcol forpolymers seems even larger (Roos and Karel,1991). This is because the decrease in viscosity atTg% may not be sufficient enough to cause struc-tural collapse (Bindschaedler, 1999). For refer-ence, Table 1 lists Tg% s and Tcols of somecommonly used excipients and buffers.

4.1.2. Crystallization temperature (Tcry)When the temperature of an aqueous protein

formulation drops below 0°C, water usually crys-tallizes out first. Then, the crystalline component,which usually has the least solubility in the formu-lation, may crystallize out. This temperature istermed crystallization temperature.

4.1.3. Eutectic crystallization/melting temperature(Teut)

When the temperature of an aqueous proteinformulation further decreases after crystallizationof the least soluble component, this componentand water crystallize out at the same time as amixture. This temperature is termed eutectic crys-tallization/melting temperature. The relationshipbetween Teut and Tg% is shown in Fig. 1. Due to

Fig. 1. A theoretical phase diagram showing ice formation,solute crystallization, eutectic point, and glass transition dur-ing freezing.


Tab

le1

Gla

sstr

ansi

tion

and

colla

pse

tem

pera

ture

s(°

C)

ofbu

ffer

s,ex

cipi

ents

and

prot

eins

Tg%

Tco

lT

gC

ompo

unds

Ref

eren

ces

Buf

feri

ngag

ents

−54

a,−

53b

11b

Cit

ric

acid

aC

hang

and

Ran

dall,

1992

;bL

uan

dZ

ogra

fi,19

97−

63C

hang

and

Ran

dall,

1992

Hep

esaC

hang

and

Ran

dall,

1992

;bO8

ster

berg

and

Wad

sten

,19

99H

isti

dine

−33

a,−

32b

Fra

nks,

1993

−76

Pot

assi

umac

etat

eF

rank

s,19

93−

62P

osta

ssiu

mci

trat

eC

hang

and

Ran

dall,

1992

−55

Pot

assi

umph

osph

ate

(KH

2P

O4)

Cha

ngan

dR

anda

ll,19

92−

64So

dium

acet

ate

Fra

nks,

1993

−52

Sodi

umbi

carb

onat

eC

hang

and

Ran

dall,

1992

−41

Sodi

umci

trat

eSo

dium

Cha

ngan

dR

anda

ll,19

92−

45ph

osph

ate

(NaH

2P

O4)

Fra

nks,

1993

Tri

sba

se−

55−

65C

hang

and

Ran

dall,

1992

Tri

s–H

Cl

Exc

ipie

nts,

low

MW

Fra

nks,

1993

b-A

lani

ne−

65R

oos,

1993

Ara

bino

se3 42

Mat

tern

etal

.,19

99A

rgin

ine

77F

rank

s,19

90;

Cos

tant

ino

etal

.,19

98b

Cel

lobi

ose

aR

oos,

1993

;bW

isni

ewsk

i,19

9810

a,

13b

Fru

ctos

e−

42b

31R

oos,

1993

Fuc

ose

−41

a31

b,

c 38G

alac

tose

aF

rank

s,19

90;

bL

evin

ean

dSl

ade,

1992

;c R

oos,

1993

−43

a−

41d

Glu

cose

4c ,32

b,

39a

aF

rank

s,19

90;

bL

evin

ean

dSl

ade,

1992

;c P

rest

rels

kiet

al.,

1995

;dA

dam

san

dR

amsa

y,19

96C

hang

and

Ran

dall,

1992

−17

Glu

tam

icac

id−

65a

−92

b,−

93a

aF

rank

s,19

90;

bB

ell

etal

.,19

95G

lyce

rol

Gly

cine

Cha

ngan

dF

isch

er,

1995

−37

b−

32a37

aM

atte

rnet

al.,

1999

;bO8

ster

berg

and

Wad

sten

,19

99H

isti

dine

−30

.5b

101a

,11

4caL

evin

ean

dSl

ade,

1992

;bA

dam

san

dR

amsa

y,19

96;

c Tay

lor

and

Zog

rafi,

1998

bL

acto

seM

atte

rnet

al.,

1999

68L

ysin

eaF

rank

s,19

90;

bL

evin

ean

dSl

ade,

1992

;c T

aylo

ran

dZ

ogra

fi,19

98b

Mal

tose

43a,

82b,

−30

a

100c

Hat

ley

and

Fra

nks,

1991

Mal

totr

iose

−23

−33

a,−

27b

13c

aM

ered

ith

etal

.,19

96;

bL

ueck

elet

al.,

1998

a;c K

imet

al.,

1998

Man

nito

l


Tab

le1

(Con

tinu

ed)

Ref

eren

ces

Tco

lC

ompo

unds

Tg

Tg%

30a,

33b

aF

rank

s,19

90;

bL

evin

ean

dSl

ade,

1992

−41

aM

anno

se91

Roo

s,19

93M

elib

iose

Wol

kers

etal

.,19

98a

Oct

ulos

e21

−26

c10

8b,

114a

aT

aylo

ran

dZ

ogra

fi,19

98b;

bW

olke

rset

al.,

1998

a;c W

illem

er,

1999

Raf

finos

e−

47b

−10

b,−

13a

aR

oos,

1993

;bW

isni

ewsk

i,19

98R

ibos

eF

rank

s,19

93So

dium

chlo

ride

B−

60aF

rank

s,19

90;

bK

err

etal

.,19

93;

c Ada

ms

and

Ram

say,

1996

−54

cSo

rbit

ol(g

luci

tol)

−3a

−44

a,b

−32

f−

31d,−

40h

aL

evin

ean

dSl

ade,

1992

;bH

anco

cket

al.,

1995

;c P

rest

rels

kiet

al.,

1995

;dA

dam

san

dSu

cros

e31

c ,61

a,

72e ,

Ram

say,

1996

;e C

osta

ntin

oet

al.,

1998

c;f K

asra

ian

etal

.,19

98;

f Lue

ckel

etal

.,19

98a;

77b,

75g

gT

aylo

ran

dZ

ogra

fi,19

98b;

hO

verc

ashi

eret

al.,

1999

aL

evin

ean

dSl

ade,

1992

;bP

rest

rels

kiet

al.,

1995

;c A

dam

san

dR

amsa

y,19

96;

dD

uddu

and

−29

c ,−

34h

78a,

46b,

117d

,T

reha

lose

115c ,

105f ,

118g

Dal

Mon

te,

1997

;e M

iller

etal

.,19

97;

f Lue

ckel

etal

.,19

98a;

gT

aylo

ran

dZ

ogra

fi,19

98b;

hO

verc

ashi

eret

al.,

1999

aB

ell

etal

.,19

95;

bM

iller

etal

.,19

97W

ater

−13

3a,−

137b

bF

rank

s,19

90;

aR

oos,

1993

−23

a,−

39b

Xyl

itol

−47

b

aF

rank

s,19

90;

bR

oos,

1993

Xyl

ose

−48

a9a

,14

b

Exc

ipie

nts,

high

MW

227

Han

cock

and

Zog

rafi,

1994

Cel

lulo

seP

rest

rels

kiet

al.,

1995

b-C

yclo

dext

rin

108

−10

b91

aD

extr

an(1

0kD

)aP

rest

rels

kiet

al.,

1995

;bW

illem

er,

1999

94a,

101b

ate

Boo

yet

al.,

1992

;bP

rest

rels

kiet

al.,

1995

Dex

tran

(40

kD)

−11

Ada

ms

and

Ram

say,

1996

Dex

tran

(70

kD)

Pik

al,

1990

b;Iz

utsu

etal

.,19

96F

icol

l−

20P

ikal

,19

90b

Gel

atin

−8

155

Han

cock

and

Zog

rafi,

1994

Hyd

roxy

prop

ylm

eth

yl-c

ellu

lose

\11

0caC

row

eet

al.,

1993

b;bP

ikal

,19

90b;

c Cro

we

etal

.,19

98−

12a

Hyd

roxy

ethy

l\

−5b

star

chM

alto

dext

rin

860

Tay

lor

and

Zog

rafi,

1998

b16

9−

9M

etho

cel

Will

emer

,19

99W

illem

er,

1999

PE

G(6

kD)

−13

Ker

ret

al.,

1993

Pol

ydex

tros

e−

27P

VP

K15

(10

kD)

aB

ell

etal

.,19

95;

bIz

utsu

etal

.,19

9610

0a−

27b

−24

Ada

ms

and

Ram

say,

1996

PV

P(4

0kD

)18

0B

ell

etal

.,19

95P

VP

K30

(40

kD)

185a

,17

6baH

anco

cket

al.,

1995

;bT

aylo

ran

dZ

ogra

fi,19

98b

PV

PK

90(1

000

kD)

Will

emer

,19

99−

10Se

phad

exG

200

Han

cock

and

Zog

rafi,

1994

227

Star

ch

Pro

tein

s−

13a,−

11b

BSA

aSl

ade

etal

.,19

89;

bC

hang

and

Ran

dall,

1992


Tab

le1

(Con

tinu

ed)

Ref

eren

ces

Tco

lC

ompo

unds

Tg

Tg%

Slad

eet

al.,

1989

−12

.5a-

Cas

ein

Slad

eet

al.,

1989

Glo

bulin

s(e

gg)

−10

.5H

SAP

ikal

,19

90b

−9.

5Sl

ade

etal

.,19

89−

10.5

a-L

acta

lbum

in(S

igm

a)C

hang

and

Ran

dall,

1992

LD

H−

9aSl

ade

etal

.,19

89;

bC

hang

and

Ran

dall,

1992

Lys

ozym

e−

16.5

a,−

13b

−20

Ove

rcas

hier

etal

.,19

99rh

uMA

bH

ER

2C

hang

and

Ran

dall,

1992

Myo

glob

in−

15aC

hang

and

Ran

dall,

1992

;bW

illem

er,

1999

−10

bO

valb

umin

−11

a

−12

Cha

ngan

dR

anda

ll,19

92R

Nas

eA


excipient interaction(s), many multicomponentprotein formulations do not exhibit Teut (Hatleyet al., 1996).

4.1.3.1. De6itrification temperature (Tde6). Whenthe temperature of the glassy maximally freeze-concentrated solution (MFCS) increases, an en-dothermic glass transition first occurs. Furtherincrease in temperature above Tg% may lead to anexothermic event, corresponding to the recrystal-lization of a component such as mannitol(Meredith et al., 1996). This temperature istermed devitrification temperature. Devitrificationis a process by which a metastable glass forms astable crystalline phase on heating above Tg%(Slade et al., 1989; Chang and Randall, 1992;Rey, 1999). Recrystallization can occur if a solu-tion has been cooled rapidly, arresting crystalnucleation and/or growth. To detect recrystalliza-tion, the heating rate should be slower than thecritical heating rate, which is defined as the mini-mum heating rate fast enough to prevent devitrifi-cation of the unfrozen fraction of a solution(Chang and Randall, 1992).

4.2. Freezing — the first step in lyophilization

The freezing step during lyophilization is con-sidered to be at least as important as the dryingstep due to its potential effect on proteins (Wille-mer, 1992). One critical parameter that needs tobe defined during freezing is the cooling rate. Thecooling rate, n, can be defined as

n=dT(r,t)

dt,

where, T(r,t) is the temperature field, a functionof both time, t, and location, r. Therefore, therate varies temporally and spatially (Hartmann etal., 1991). In general, a faster freezing rate gener-ates small ice crystals (Eckhardt et al., 1991;Willemer, 1992; Wisniewski, 1998). This is be-cause water is super-cooled and crystallizationinto ice occurs rapidly, producing small ice crys-tals (Pikal, 1990a; House and Mariner, 1996).Conversely, a slower cooling rate generates larger

ice crystals. The size of crystals determines thepore size to be created during subsequent drying.Large ice crystals create large pores, leading torapid water sublimation during primary drying(Willemer, 1992), but the secondary drying mayslow down due to smaller surface areas, limitingwater desorption during secondary drying (Bind-schaedler, 1999). To keep a balance, a moderatedegree of supercooling (10–15°C) has been rec-ommended (Pikal, 1990a).

The rate of freezing-induced protein denatura-tion is a complex function of both cooling rateand final temperature (Franks, 1990). The effectof cooling rates on the stability of proteins variessignificantly. For example, increasing the freezingrate from 0.5 to 50°C min−1 did not significantlyaffect the formation of soluble aggregates of rGHat 2 mg ml−1 in 5 mM phosphate buffer (pH 7.4or 7.8), but the formation of insoluble aggregates(particulates) increased sharply with increasingcooling rates, even in the presence of up to 250mM mannitol (Eckhardt et al., 1991). Rapidfreezing also caused formation of more aggregatesfor bovine and human IgG than slow freezing(Sarciaux et al., 1998). This may result from theformation of smaller ice crystals and larger ice-water interfaces at higher freezing rates, leadingto a greater extent of surface-induced proteindenaturation. On the contrary, faster freezingcaused less loss of LDH activity (Nema and Avis,1992) and less change in the secondary structureof hemoglobin in a PEG/dextran solution (Helleret al., 1999a). The lower loss of protein activity islikely because faster freezing may prevent exten-sive crystal growth, which may substantially hin-der solute concentration-induced protein dena-turation. Therefore, stability of proteins may beaffected differently at different freezing rates de-pending on protein denaturation mechanisms.

It should be noted that freezing rate may havea potential impact on the storage stability oflyophilized proteins. Hsu et al. (1995) demon-strated that faster cooling during lyophilization oftPA resulted in a product cake with larger inter-nal surface area, which led to formation of moreopalescent (insoluble) particulates upon long-term


storage at 50°C. The rate of formation of opales-cent particulates during storage correlated well(r=0.995) with internal surface area of thelyophilized tPA product cake.

As discussed in Section 2.2, buffer species maycrystallize out selectively and cause pH shiftingduring freezing. Therefore, it is preferable to keepall buffer species amorphous during freezing ofpH-sensitive proteins. A faster cooling rate mayprevent nucleation and subsequent crystallization(Franks, 1993). Each buffer salt has its own criti-cal cooling rate, which is defined as the minimumcooling rate fast enough to prevent crystallizationof a solute (Chang and Randall, 1992). Crystal-lization will not occur when the cooling rate ishigher than the critical cooling rate. If a proteinsolution is cooled rapidly with liquid nitrogen, itis likely that buffer salts will remain amorphous.Otherwise, selective crystallization and a subse-quent pH change may occur (Chang et al.,1996b).

Freezing rate influences the extent of crystal-lization of a formulation excipient, such as manni-tol (Hsu et al., 1996). Accordingly, the duration ofsubsequent thermal treatment after freezing canbe affected (see next section). In addition, differ-ent freezing rates may favor formation of certaincrystalline forms of an excipient, which may po-tentially affect protein stability and reconstitution.It has been observed that a slower rate of crystal-lization tends to favor formation of g-glycine,whereas rapid crystallization seems to favor for-mation of the b-polymorph (Akers et al., 1995).Different polymorphs of mannitol were also ob-tained at different concentrations during freezing(Izutsu et al., 1993). Recently, Kim et al. (1998)demonstrated that slow freezing (about 0.2°Cmin−1) of 10% (w/v) mannitol produced a mix-ture of a and b-polymorphs and fast freezing (byliquid nitrogen) of the same solution producedd-form. The reconstitution time (with water) was36 and 78 s, respectively, for the fast-freeze andslow-freeze dried mannitol samples.

What is the ideal location in a product vialwhere the cooling rate should be measured? Hart-mann et al. (1991) demonstrated that the coolingrate in a flat plate-shaped freezing container, ver-tically submerged into liquid nitrogen, generally

became higher from the surface to the center ofthe container. They found that the optimum loca-tion where the cooling rate should be measured,was 1/3 away from the center and 2/3 away fromthe inside surface of the sample container. Thecooling rate at this point represented \80% ofthe entire sample volume. The geometrical centerwas apparently the worst location (the least repre-sentative) for measuring cooling rate. Althoughthe optimum location for measuring cooling ratein a cylindrical vial has not been determined, itmay not be at the center based on the abovestudy.

4.3. Thermal treatment prior to drying

Very often, a thermal treatment step, annealing,is included before the primary drying step. Thereare at least two reasons for this. First, during thefreezing step, a crystalline component may not becompletely crystallized. Complete crystallizationmay be necessary if this component is to providenecessary cake structure or if the protein is morestable after complete crystallization. Recrystalliza-tion can be promoted by heat treatment at tem-peratures above Tg% of the formulation. Theduration of the treatment depends on the compo-sition of the formulation and the heating ratethrough Tg% (Getlin, 1991). Second, removal fromthe amorphous phase of a crystalline componentwhich has a low Tg% , such as glycine (Tg%= −42°C), can increase the Tg% of the amorphousphase (Carpenter et al., 1997). The increased Tg%can allow more efficient primary drying at ahigher temperature. For example, annealing at−20°C in a glycine:sucrose (1:1 weight ratio)formulation increased the Tg% from −44 to −33°C and produced a formulation cake of betterappearance and higher mechanical strength(Lueckel et al., 1998a,b). Similarly, incorporationof an annealing step in the lyophilization of amonoclonal antibody crystallized glycine and en-abled a higher drying temperature (Ma et al.,1998).

However, an annealing step may have an ad-verse effect on the stability of a protein due tocrystallization of an amorphous stabilizer, losinghydrogen bond interaction with the protein (also


see Sections 3.2 and 5.2). Such cases occurred toseveral proteins during lyophilization, includingb-galactosidase (2 mg ml−1) in mannitol formula-tion during lyophilization (Izutsu et al., 1993),interleukin-6 (IL-6) in sucrose:glycine (1:1 at 20mg ml−1) formulation (Lueckel et al., 1998b), andLDH in mannitol formulation (Izutsu et al.,1994b). Annealing at −7°C for 1 or 12 h was alsofound to destabilize hemoglobin in PEG/dextrin(1:1 weight ratio) system during lyophilization, asmonitored by IR (Heller et al., 1999c).

4.4. Drying

The drying step is divided into two phases:primary and secondary drying. The primary dry-ing removes the frozen water (sublimation of ice)and the secondary drying removes non-frozen‘bound’ water (desorption of water). The amountof non-frozen water for globular proteins is about0.3–0.35 g g−1 protein, slightly less than theproteins’ hydration shell (Rupley and Careri,1991; Kuhlman et al., 1997).

Different models have been reported to describethe drying/sublimation rate during lyophilization(Jennings, 1999). When the shelf temperature isfixed, the drying/sublimation rate, n, of a frozensolid can be expressed as:

n=Ap(Pp−P0)

Rp

in which Ap is the cross sectional area of aproduct, Pp is product vapor pressure at the subli-mation front, P0 is partial vapor pressure in aproduct vial, and Rp is resistance of a driedproduct layer to vapor flow (Nail and Johnson,1992). Rp may be different under different freezingconditions. Since Ap, Pp, and Rp at a fixed freez-ing rate are usually fixed for a particular proteinformulation in a chosen container, drying rate canbe changed only through adjustment of P0.

As indicated in the above equation, drying rateis inversely proportional to Rp. If Rp changesduring the drying process, the drying rate and theproduct temperature will change accordingly.During lyophilization, continuous drying maylead to formation of an increasingly dry productlayer (Overcashier et al., 1999). This formulation-

dependent dry layer hinders diffusion of watervapor, increases Rp, and causes product tempera-ture to rise, as observed for Erwinia L-asparagi-nase during lyophilization (Adams and Ramsay,1996). Any temperature rise during lyophilizationmay potentially cause product collapse. In thiscase, the shelf temperature needs to be reducedaccordingly to prevent product collapse or melt.On the other hand, small-scale product collapseduring lyophilization may decrease Rp (Over-cashier et al., 1999).

Several other factors, such as ice morphology,crystal size distribution (see Section 4.2), and for-mulation composition, also affect the drying rateor time. Excipients in protein formulation mayinteract with proteins and reduce the availabilityof water-binding sites in proteins. At the sametime, excipients themselves may interact with wa-ter molecules. Thus, the water-binding force ofthe formulation components and the associatedamount of monolayer water molecules coveringthe protein formulation will be different depend-ing on the formulation composition, as demon-strated in the rhGH:sugar or rhIGF-I:sugarformulations (Costantino et al., 1998c). Conse-quently, the drying time will be different depend-ing on formulation composition. In addition,composition-dependent formation of any excipi-ent hydrates during lyophilization would invari-ably reduce the drying rate, as implicated in theformation of a mannitol hydrate duringlyophilization (Yu et al., 1999).

The driving force for water sublimation duringlyophilization is the temperature difference be-tween the product and the condenser. The com-monly used condenser temperature is −60°C,allowing a minimum of 20°C lower than theproduct temperature during primary drying(Franks, 1990). During secondary drying the con-denser temperature can be set even lower, such as−80°C, for formulations that require very lowresidual moisture (Bindschaedler, 1999). Toachieve a high drying rate, product temperature isfrequently set as high as possible. Since theproduct temperature is controlled by the shelftemperature, effective heat transfer between shelfand product is essential, which is affected by thedegree of vacuum in the drying chamber. A mod-


erate increase in chamber pressure often increasethe drying rate due to more effective heat transfer,leading to a higher product temperature (Bind-schaedler, 1999). In the development of alyophilized vaccine formulation, the convection ofheat between shelf and product was shown to bemore effective under a vacuum of 100–120 mTorrthan under maximum vacuum (15–25 mTorr)(House and Mariner, 1996). For this reason, dry-ing of the formulation was not conducted undermaximum vacuum. Similarly, it has been reportedthat the maximum drying rate at a chamber pres-sure of 400 mmHg is more than twice that at 100mmHg at a shelf temperature of 40°C (Nail andJohnson, 1992). In a more recent report, it wasshown that at a constant shelf temperature of25°C, the specific sublimation rate was 0.19, 0.16,and 0.11 g h−1 cm−2 at chamber pressures of300, 200, and 100 mTorr, respectively, for aprotein formulation containing trehalose, his-tidine, and polysorbate 20 (Overcashier et al.,1999). Therefore, a balanced degree of vacuum inthe drying chamber is needed to achieve the de-sired drying rate. It has been suggested that achamber pressure at one-fourth to one-half of thesaturated vapor pressure over ice usually lead to ahigh sublimation rate (Bindschaedler, 1999).

During lyophilization, complete sample collapseresults in both a lower rate of water sublimation/desorption and an inferior product. In addition,collapsed materials may crystallize more easilythan non-collapsed materials during storage(Darcy and Buckton, 1997). To prevent productcollapse, the product temperature must be keptbelow the glass transition temperature (Tg% ) (orTcol) of the formulation or below the eutecticmelting temperature (Teut) of any crystalline com-ponent. On the other hand, primary drying shouldbe operated at a temperature as close as possibleto these temperature limits for high efficiency.Therefore, to have an efficient drying step and toreduce probability of product collapse, a formula-tion should be designed such that its Tg% is as highas possible. To ensure a high Tg% during secondarydrying, the primary drying cycle should be com-pletely finished so that only a minimum amountof bound water is left in the formulation. Thereare other advantages of drying at a relatively high

temperature. Sheehan and Liapis (1998) recentlymodeled primary and secondary drying of phar-maceutical products in vials. They found thatcontrolling heat input close to melting and scorch(thermally damaging) temperature constraints re-sulted in not only faster drying time but also moreuniform distribution of temperature and boundwater in the formulation at the end of secondarydrying. The scorch temperature was defined as thetemperature of the top surface that shows thermaldamage and the melting temperature of the frozenphase was considered to be about 10°C below themelting point of the ice.

The end of the primary drying process is whenall the frozen water is removed and the rate ofwater sublimation is significantly reduced. Severalmethods can be used in monitoring the comple-tion of the primary drying cycle. A simple methodis to observe the changes in product temperature(or chamber pressure) during freeze-drying. Theend of the primary drying process is when theproduct temperature approaches the shelf temper-ature, evidenced by a significant change in theslope of the product temperature trace due to areduced sublimation rate. To prevent prematureending, an extra 2–3 h may be added to thedrying cycle. A more objective method is thepressure rise test. By disconnecting the vacuumsource, the chamber pressure should rise at a ratedepending on the amount of moisture in theproduct. The end of the drying process would bewhen the rate of pressure rise is below a specifiedvalue. Another method for determining the end ofthe primary drying process is the measurement ofthe heat transfer rate (Jennings and Duan, 1995;Oetjen, 1999). Timely detection of the end of theprimary drying cycle by this method have resultedin a more efficient drying process (Jennings andDuan, 1995). The duration of the secondary dry-ing cycle is dictated by the required moisturecontent in the final product (see next section).

To improve drying efficiency, a single-dryingstep may be designed for certain proteins. Changand Fischer (1995) developed such a cycle forrhIL-1ra formulation. In this single cycle, theshelf temperature was set for the secondary dryingand the product temperature was controlled to acertain level by adjusting the chamber tempera-


ture. Freeze-drying of 1 ml of rhIL-1ra formula-tion could be completed within 6 h with a finalmoisture level of 0.4%. However, the cycle-associ-ated heterogeneity in temperature and moisturedistribution within the product may cause poten-tial damage to other labile proteins.

4.5. Residual moisture

The desired residual moisture level in alyophilized product dictates duration of the sec-ondary drying step. An electronic hygrometer or aresidual gas analyzer may be used to determineresidual moisture level during lyophilization andthus, the end-point of secondary drying (Nail andJohnson, 1992). The aforementioned pressure risetest or the measurement of heat transfer rate canalso be used for determination of the end of thesecondary drying cycle. If these methods are notavailable, the minimum duration of drying mayhave to be determined systematically using differ-ent combinations of shelf temperatures and dura-tions (Greiff, 1992). Moisture content oflyophilized formulations can be determined byseveral methods, including loss-on-drying, KarlFischer titration, thermal gravimetric analysis(TGA), gas chromatography (GC), or near IR.The reproducibility of moisture determination bythe first three methods was found to be similar forseveral biological products (May et al., 1989).

What is the residual moisture in lyophilizedproduct? Proteins have both strong and weakbinding sites to accommodate unfrozen water.The weak binding sites include mostly carbonylbackbone plus hydrophilic –OH and –NH–groups, while the strong binding sites includethose ionizable groups in amino acids such asGlu, Asp, Lys, and Arg (Careri et al., 1979). Forlysozyme, strongly bound water moleculesamount up to 10% (g g−1 protein) at 38°C (Careriet al., 1979). Since the residual moisture contentfor lyophilized protein products is usually below10%, secondary drying removes weakly boundand some of strongly bound water molecules.Therefore, the residual moisture is a small portionof strongly bound water molecules in proteins.

Residual moisture causes a variety of instabili-ties in dried proteins and in many cases, the effect

is complex (see Section 5.2). Usually, a lowermoisture content leads to a more stable proteinproduct, although there may not be any signifi-cant difference in protein stability between near-zero and an intermediate moisture content ofabout 1% (Pikal, 1990a). On the other hand, if alyophilized formulation needs additional viral in-activation by dry heat, its moisture content needsto be high enough to achieve efficient and effec-tive inactivation (Savage et al., 1998). As a gen-eral rule, a moisture content in a lyophilizedprotein formulation should not exceed 2%(Daukas and Trappler, 1998). This general ruleseems applicable at least to a couple of proteinformulations. Lyophilized bFGF with sugars wasstable as long as the moisture content was below2% (Wu et al., 1998). Lyophilized monoclonalantibody cA2 IgG remained stable with a mois-ture content of 2.2% or less (Katakam et al.,1998). However, lyophilized BSA and bovine g-globulin (BGG) formulations were more stable ata water content of about 10% than at B1%(Yoshioka et al., 1997). The complicated effect ofwater on stability of solid protein formulations isdiscussed in Section 5.2.

To find the moisture content that confers themaximal stability for a lyophilized proteinproduct, long-term stability studies should be con-ducted on protein formulations with differentmoisture contents. Only these real-time stabilitystudies can determine the optimal moisture con-tent for the final protein product.

5. Instability, stabilization, and formulation ofsolid protein pharmaceuticals

Lyophilized proteins may lose activity rapidlyduring storage, even though they may be stableduring lyophilization. For instance, porcine pan-creatic elastase without excipients retained its fullactivity after freeze-drying, but lost about 70%activity in 2 weeks at 40°C and 79% RH (Changet al., 1993). Therefore, lyophilized proteins stillneed stabilization in the solid state to survivelong-term storage as pharmaceuticals. In the fol-lowing section, instability pathways of proteins insolid state during storage are first described in


brief, followed by a detailed description of factorsaffecting stability of solid proteins and variousstabilization strategies. In the final section, severalaspects are discussed in relation to formulation ofacceptable solid protein pharmaceuticals.

5.1. Instability of solid proteins during storage

A variety of instability mechanisms have beenreported of lyophilized proteins during storage.These include aggregation (a major physical insta-bility) and different chemical degradations such asdeamidation, browning reaction, oxidation, hy-drolysis, and disulfide bond formation/exchange.

5.1.1. Protein aggregationAggregation is one of the major instabilities for

lyophilized protein pharmaceuticals during stor-age (Costantino et al., 1998d). Many lyophilizedproteins form aggregates easily during storageunder accelerated conditions, such as BSA at 37or 60°C (Liu et al., 1990; Jordan et al., 1994),rHA at 37°C and 96% RH (Costantino et al.,1995a,b), porcine pancreatic elastase at 40°C and79% RH (Chang et al., 1993), Humicola lanugi-nosa lipase at 40 or 60°C (Kreilgaard et al., 1999),IL-2 at 45°C (Zhang et al., 1996) or 65°C (Kenneyet al., 1986), rhIL-1ra at 50°C (Chang et al.,1996a), insulin at 50°C and 96% RH (Costantinoet al., 1994b), tPA at 50°C (Hsu et al., 1995), andtumor necrosis factor (TNF) at 37°C (Hora et al.,1992a). Some proteins form aggregates duringstorage under ambient conditions. These includeaFGF (Volkin and Middaugh, 1996), b-galactosi-dase (Yoshioka et al., 1993), hGH (Pikal et al.,1992; Costantino et al., 1998b), and antibody-vinca conjugate (Roy et al., 1992).

Protein aggregation can be physical, chemical,or both. Physical (non-covalent) interaction is thecause of protein aggregation for tetanus toxoid(Costantino et al., 1994a), ovalbumin (chicken eggalbumin), and glucose oxidase (Liu et al., 1990).Disulfide bond formation or exchange is a majorchemical (covalent) pathway leading to proteinaggregation. Proteins that aggregate by this mech-anism include b-galactosidase (Yoshioka et al.,1993), insulin (Costantino et al., 1994b; Strickleyand Anderson, 1996; Pikal and Rigsbee, 1997),

rHA (Costantino et al., 1995b), BSA and b-lac-toglobulin (Liu et al., 1990; Jordan et al., 1994).Both covalent (disulfide bond) and non-covalentformation of dimers and trimers were observedfor a spray-dried anti-IgE monoclonal antibody(Andya et al., 1999).

Often, both soluble and insoluble protein aggre-gates can form at the same time during storage.This is the case for hGH (Pikal et al., 1992;Costantino et al., 1998b), Humicola lanuginosalipase (Kreilgaard et al., 1999), and tPA (Hsu etal., 1995). The relative amounts of soluble andinsoluble protein aggregates may change withstorage conditions such as lysozyme aggregationunder different relative humidities (Separovicet al., 1998). Both physical and chemical aggrega-tion can lead to formation of insoluble aggregatessuch as moisture-induced aggregation oflyophilized insulin during storage (Costantino etal., 1994b).

5.1.2. Chemical degradationsGenerally speaking, chemical degradations of

proteins in solid state have not been reported asextensively as in liquid state. Nevertheless, severalchemical degradation pathways have been ob-served in lyophilized proteins during storage. Insome cases, multiple degradation processes pro-ceed simultaneously in a protein such aslyophilized rGH, which undergoes methionine ox-idation, asparagine deamidation, and irreversibleaggregation during storage (Pikal et al., 1992).Detailed mechanisms of chemical degradations insolid proteins and peptides have recently beenreviewed (Lai and Topp, 1999).

Chemical degradations may not affect the activ-ity of proteins, depending on the location of thetransformed residue(s). Due to the terminal loca-tion of MetB4 and MetB25 in recombinant humanrelaxin, oxidation of these two residues did notchange the protein bioactivity (Nguyen et al.,1993). The Met1 mono-oxidized recombinant hu-man leptin (16 kD) did not show any detectablechanges in tertiary structure and retained its fullpotency, as compared with the native form (Liu etal., 1998). Other protein degradation productshaving essentially the same biological activity asthe intact proteins include deamidated insulin


(Brange et al., 1992b), oxidized IL-2 (Kenney etal., 1986), and deamidated rIL-2 (Sasaoki et al.,1992).

5.1.3. DeamidationDeamidation is one of the major degradation

pathways in lyophilized proteins during storage.Asn and Gln are the two amino acids susceptibleto deamidation in proteins. Many cases of deami-dation have been reported in lyophilized proteins.For example, both lyophilized rGH and bFGFdeamidated during storage (Pikal et al., 1992; Wuet al., 1998). Insulin, lyophilized from a solution ofpH 3–5, deamidated via a cyclic anhydride inter-mediate at C-terminal AsnA21 in addition to cova-lent dimerization during storage (Strickley andAnderson, 1996). The combined formation ofdeamidated insulin and insulin dimers wasshown to be a linear function of square root oftime (Pikal and Rigsbee, 1997). Storing lyophilizedIL-1ra (in 2% glycine, 1% sucrose, and 10 mMsodium citrate buffer) at 50°C caused proteindeamidation in addition to aggregation (Chang etal., 1996a,c).

5.1.4. Maillard reactionReducing sugars such as glucose can react with

lysine and arginine residues in proteins to formcarbohydrate adduct via the Maillard reaction,which is also called the browning reaction (Paulsenand Pflughaupt, 1980). The Maillard reaction hasbeen a subject of extensive investigation mainly inthe food industry (Chuyen, 1998). In the develop-ment of solid protein pharmaceuticals, this reac-tion has also been observed in several lyophilizedproteins during storage, including aFGF (Volkinand Middaugh, 1996), bFGF (Wu et al., 1998),human relaxin (Li et al., 1996), IgG (Hekman etal., 1995), and porcine pancreatic elastase (Changet al., 1993). This glycation reaction also occurredto an anti-IgE monoclonal antibody co-spray-dried with lactose during storage (Andya et al.,1999).

The browning reaction can result in significantinactivation of a lyophilized protein during stor-age. Storing lyophilized invertase in the presenceof raffinose, lactose, or maltose at 95°C for 7 daysled to significant browning and the color intensity

of the reconstituted protein solution correlatedpositively with the activity loss of the protein(Schebor et al., 1997). The lactose or maltose-in-duced browning reaction may explain why thesetwo sugars were less effective than sucrose, anon-reducing sugar, in stabilizing vacuum-driedrestriction enzyme EcoRI during storage at 45°C(Rossi et al., 1997).

Although sucrose is a non-reducing sugar, it canbe easily hydrolyzed into two reducing sugars,especially at low pHs during storage, not only inliquid state (Reyes et al., 1982; Buera et al., 1987)but also in solid state (te Booy et al., 1992;Skrabanja et al., 1994). It has been demonstratedthat the rate of color formation in a freeze-driedsucrose/lysine formulation at pH 2.5 at 40°C wasclose to that in a glucose/lysine formulation, partlydue to the catalytic effect of the amino acid onsucrose hydrolysis (O’Brien, 1996). Therefore,lyophilized proteins in a sucrose-containing formu-lation may still have the potential to experience thebrowning reaction. This is the case for lyophilizedIL-6 during storage (Lueckel et al., 1998b).

5.1.5. OxidationThe side chains of Met, Cys, His, Trp, and Tyr

residues are potential sites of oxidation (Manninget al., 1989). Methionine residues in proteins caneasily be oxidized by atmospheric oxygen.Lyophilized hGH was easily oxidized in a vialcontaining only 0.4% oxygen during storage at25°C (Pikal et al., 1991, 1992). The methionineoxidation in hGH during storage has been provedto be insensitive to moisture. Other examples ofmethionine oxidation in proteins includelyophilized hIGF-I (Fransson et al., 1996) andIL-2 (Kenney et al., 1986; Hora et al., 1992b).

The free sulphydryl groups of cysteines in solidproteins can also be easily oxidized to formdisulfide bridges during storage. This is one of themajor chemical mechanisms of covalent proteinaggregation as mentioned above.

5.1.6. HydrolysisAlthough moisture content is usually low in

lyophilized protein formulations, hydrolysis canstill occur during storage. Hydrolysis of bFGF hasbeen observed recently in a lyophilized sugar


formulation (Wu et al., 1998). Li et al. (1996)found that the loss of relaxin activity (48%) in alyophilized glucose formulation during storage at40°C for 2 weeks was significantly higher thanthat (8%) in either mannitol or trehalose formula-tion. The loss of protein activity was apparentlydue to glucose-induced elimination of serine at theC-terminal of the B chain in relaxin to formdes-Ser relaxin.

5.2. Factors affecting stability of solid proteins

Several factors can affect the stability of solidprotein pharmaceuticals. These include storagetemperature, glass transition temperature, formu-lation pH, residual moisture content, type andconcentration of formulation excipients, and crys-tallization of amorphous excipients.

5.2.1. Storage temperatureStorage temperature is probably one of the

most important factors affecting protein stabilityin solid state. Many lyophilized proteins showincreased loss of activity at high temperatures,such as restriction enzyme PstI in a trehaloseformulation and neutral lactase in a PVP formula-tion at temperatures between 37 and 70°C (Co-laco et al., 1992; Mazzobre et al., 1997).Unfortunately, due to the structural complexitiesof proteins, the temperature effect on stability ofsolid proteins cannot be simply described by asingle instability mechanism, although in general,the higher the temperature, the lower the proteinstability, both physically and chemically. It shouldbe noted that fluctuating storage temperaturesmay be more detrimental to a lyophilized proteinthan a single high storage temperature (Ford andDawson, 1994).

High temperatures accelerate physical aggrega-tion of proteins in solid state. This can be ascribedto an increased mobility of protein molecules athigh temperatures, which facilitates protein–protein interactions. Increasing temperatures in-creased aggregation of lyophilized rhIL-1rabetween 8 and 50°C (Chang et al., 1996a,c), ag-gregation of vacuum-dried LDH between 3 and60°C, and aggregation of vacuum-dried rhG–CSF between 3 and 80°C (Mattern et al., 1997,

1999). High temperatures also accelerate chemicaldegradations of proteins in solid state, such asdeamidation of rhIL-1ra at temperatures between8 and 50°C (Chang et al., 1996a,c) and dimeriza-tion of TNF at temperatures between 25 and 80°C(Hora et al., 1992a).

Increasing temperature also affects protein sta-bility indirectly. Since proteins in a solid formula-tion are stabilized in an amorphous phase,crystallization of an amorphous component(s)may destabilize proteins. Crystallization of anamorphous component in a protein formulationcan occur slowly during storage below Tg andrapidly above Tg (Sun, 1997). The effect of crys-tallization on protein stability will be discussed ina following section.

5.2.2. Glass transition temperature (Tg)Glass transition temperature (Tg) of protein

formulations is considered to be one of the majordeterminants of protein stability (Hatley andFranks, 1991). Tg of a polymer is defined as thetransition temperature between the rubbery (orliquid-like) and glassy (solid-like) states. Micro-scopically, a polymer chain may have cooperativelocalized motion above Tg, and below Tg, onlyindividual atoms are able to make small excur-sions about their equilibrium positions (Ravve,1967). The greater mobility above Tg is due to agreater free volume and higher degree of bothtranslational and rotational freedom (Slade et al.,1989). Therefore, formation of a glassy stateshould result in a drop in molecular diffusion/mo-tion and thus, increased protein stability (Dudduand Dal Monte, 1997).

DSC is the most widely used method for deter-mination of Tg (Crowe et al., 1998). However, theTg of a protein formulation may not be easilydetectable or accurately determined by DSC dueto its poor reproducibility, limited detectability,formulation heterogeneity, and contribution ofsecondary (or b-) relaxation processes (Fan et al.,1994; Bell et al., 1995; Pikal and Rigsbee, 1997;Yoshioka et al., 1997, 1998). Several other factorscan affect Tg and/or its determination, such as thecondition of glass formation, moisture content,temperature ramping rate, thermal history of sam-ples, or presence of multiple temperature transi-


tions (Franks, 1994; Her and Nail, 1994; Sartor etal., 1994; Shamblin et al., 1998; Lueckel et al.,1998a). These factors may partly explain why Tgsare so different for the same compound, as re-ported by different investigators (Table 1).

Generally, the higher the glass transition tem-perature, the more stable the protein formulation.Therefore, the glass transition temperature maybe used with caution as a guiding parameter toscreen protein stabilizers and formulations. Isthere a minimum Tg, which can considered to beadequate to achieve long-term stability for solidprotein pharmaceuticals? It has been recom-mended that the Tg of a stable protein productshould be at least 20°C above ambient storagetemperature (Franks, 1994). In other words, theproduct Tg should be higher than 40°C to achievelong-term stability and to tolerate shippingstresses (Carpenter et al., 1997). When Tg is abovethis temperature, protein molecular mobility isrestricted, thus minimizing protein reactivity(Pikal et al., 1991).

Is there any significant difference in molecularmobility immediately below and above Tg, whichleads to a significant difference in protein stabil-ity? Oksanen and Zografi (1993) found that therewas no critical (sharp) change in diffusion coeffi-cient (mobility) of water in PVP around Tg andsignificant mobility existed below Tg. Significantmolecular mobility was also found of in-domethacin, PVP, and sucrose at temperaturesimmediately below Tg (Hancock et al., 1995).Therefore, significant protein degradation doesoccur below Tg, such as the degradation oflyophilized human insulin via formation of acyclic anhydride intermediate at C-terminalAsnA21 (Strickley and Anderson, 1996). In a re-cent study, thermal inactivation of lactase in alyophilized PVP formulation was investigated attemperatures between 37 and 70°C and the inacti-vation rate constant did not show a step changearound the Tg (55°C) of the formulation (Mazzo-bre et al., 1997).

How well can Tg be used to predict stability ofsolid protein pharmaceuticals? Mixed results havebeen obtained. The storage stability of freeze-dried IL-2 correlated positively with Tg of theformulations in the presence of stabilizers of dif-

ferent molecular weights, including glucose (180D), sucrose (342 D), trehalose (342 D), raffinose(505 D), stachyose (667 D), b-cyclodextrin (1135D), dextran 12 (10 kD), and dextran 40 (39 kD)(Prestrelski et al., 1995). Positive correlation wasalso observed between the stability of freeze-driedinvertase at 90°C and the glass transition temper-ature of polymer excipients including maltodex-trin and PVPs (PVP 10, 40, and 360) (Schebor etal., 1996). The denaturation temperature (Tm) ofseveral dried proteins, including b-lactoglobulin,ovalbumin, lysozyme, somatotropin, and ribonu-clease A, correlated positively with Tg of the sugaror polyol additives (Bell et al., 1995; Bell andHageman, 1996). The thermal stability of proteinsin seed axes strongly depends on Tg of the intra-cellular glass (Sun et al., 1998). However, thereare many examples where formulations of a lowerTg are more stable than those of a higher Tg,especially when sugar formulations are comparedwith polymer formulations. For example, the Tg

values of both sucrose and trehalose are signifi-cantly lower than those of maltodextrin and PVP(40 kD) but both sugars were more effective thanthe polymers in stabilizing vacuum-dried restric-tion enzyme EcoRI during storage at 37 or 45°C(Rossi et al., 1997). Similarly, the Tg of freeze-dried invertase formulations containing trehalose,maltodextrin, or PVP was 63, 124, and 119°C,respectively, and the remaining activity of inver-tase in the respective formulations after incuba-tion at 90°C for 6 h was 74, 47, and 56%(Cardona et al., 1997). The inferiority of polymersin stabilizing proteins has been attributed to theirinefficient hydrogen bonding with proteins (seeSection 3.1). In a different study, Chang et al.(1996a) found that the Tg of a lyophilized rhIL-1ra formulation containing either phosphate orcitrate was 26 and 46°C, respectively, but bothdeamidation and aggregation of the protein inphosphate-containing formulation were signifi-cantly slower during storage at 0°C. They alsofound that increasing the sucrose concentrationfrom 1 to 10% (w/v) in rhIL-1ra formulation (2%glycine and 10 mM citrate, pH 6.5) decreased theTg of the lyophilized formulation from 68.5 to64.6°C, but both aggregation and deamidation ofthe protein were significantly slower during stor-age at 50°C.


Why does Tg sometimes fail to predict proteinstability? The reason is that Tg is a critical temper-ature of molecular mobility of amorphous materi-als, but may not be a direct indicator of molecularmobility (Yoshioka et al., 1998). For instance, theTg of a sucrose-containing chimeric monoclonalantibody formulation was lower than that of atrehalose-containing formulation, but the molecu-lar mobility in the sucrose-containing protein for-mulation was apparently lower than that in thetrehalose-containing formulation at low tempera-tures (B12°C), possibly due to the differences inglass fragilities of the two formulations (Duddu etal., 1997). Probably due to the lower molecularmobility, protein aggregation in the sucrose for-mulation during storage at 5°C was slightly lowerthan that in the trehalose formulation. Therefore,Tg may not be a direct indicator of protein mobil-ity. Instead, the Kauzmann temperature (TK),which is the extrapolated isoentropy temperaturebelow both Tm and Tg, has been suggested asbeing more of a molecular mobility indicator(Hancock and Zografi, 1997; Hancock et al.,1998). This is because TK is frequently coincidedwith the zero mobility temperature (T0) (Angell,1995).

Recent studies indicate that Tmc appears moreclosely related to protein stability than Tg. Tmc

was defined as the molecular mobility-changingtemperature at which protein or excipient protonsin lyophilized formulations begin to exhibitLorentzian relaxation process resulting fromhigher molecular mobility in addition to Gaussianrelaxation process resulting from lower molecularmobility (Yoshioka et al., 1997, 1998, 1999). AtTmc, protein formulations transition from a non-liquidized state to a microscopically liquidizedstate. Tmc has been found to be more closelyrelated to stability of lyophilized g-globulin thanTg (Yoshioka et al., 1997, 1998). The molecularmobility of proteins can be determined by NMRor dielectric analysis (Pearson and Smith, 1998).

5.2.3. Formulation pHThe pH of a protein solution for lyophilization

often affects the stability of dried protein productsduring long-term storage. The solution pH of asolid formulation upon reconstitution is consid-

ered to be a measure of the solid-state microenvi-ronment ‘pH’ (Strickley and Anderson, 1997).Therefore, solid-state acidity/basicity may still af-fect protein stability, both physically and chemi-cally. For example, the formation ofnon-dissociable aggregates during storinglyophilized RNase depended on the pH of theprotein solution for lyophilization in the followingorder, pH 10.0\4.0\6.4\ water (Townsendand DeLuca, 1990). At a storage temperature of25°C, aggregation of a lyophilized antibody-vincaconjugate from a solution of pH 8.5 was morethan that from pH 7.1 or 6.1 (Roy et al., 1992).Both the remaining activity and dimer formationof lyophilized TNF during storage varied withsolution pHs from 4.0 to 10.0 (Hora et al., 1992a).Some lyophilized proteins, however, are insensi-tive to pH changes. Lyophilization of ovalbuminfrom solutions of different pHs (3.0, 5.5, 7.3, and9.0) had little effect on the formation of non-co-valent protein aggregates during storage (Liu etal., 1990).

In a few cases, the acidity/basicity of a solidformulation affects chemical stability of a protein.The rate and product distribution (between cova-lent dimer and deamidated form) of insulin degra-dation in a lyophilized formulation depended onthe pH of the reconstituted protein solution(Strickley and Anderson, 1996, 1997). The rate ofdeamidation of lyophilized IL-1ra (2% glycine and1% sucrose in 10 mM sodium citrate buffer) at50°C depended on the solution pH in the follow-ing order, pH 7.0\6.5\6.0\5.5 (Chang et al.,1996c).

One of the major chemical aggregation mecha-nisms is formation of intermolecular disulfidebonds, which requires the presence of a free thio-late ion, the reactive group. A free thiolate ioncan be brought about upon ionization of a freethiol group under alkaline conditions or via b-elimination of an intact disulfide, a reaction thatis also accelerated under alkaline conditions.Therefore, the solution pH significantly affectsprotein aggregation through these mechanisms.For example, insulin lyophilized at 1 mg ml−1 atpH 10 aggregated completely in just 1 day at 50°Cand 96% RH, whereas insulin lyophilized at pH7.3 exhibited aggregation of less than 50% in 3


weeks under the same conditions (Costantino etal., 1994b). rHA lyophilized at pH 4.0 was muchmore stable than that lyophilized at pH 7.3 or 9.0against moisture-induced aggregation (Costantinoet al., 1995a).

The ‘pH’ of a solid protein formulation also hasindirect effects on protein stability. The pH-in-duced change in the hygroscopicity of a protein issuch an effect. At relatively low relative humidity(575%), the equilibrated moisture content of alyophilized insulin formulation is similar at pH3.1 and 5.0 at 35°C, but at higher relative humid-ity, the low-pH formulation is much more hygro-scopic than that of high pH (Strickley andAnderson, 1996). Sucrose at lower pHs may hy-drolyze readily to form fructose and glucose,which can react with proteins via the Maillardreaction (see Section 5.1). It was found that theactual sucrose content of freeze-dried sucrosefrom a solution of pH 3 (in citrate and phosphatebuffer) dropped to 79% at 4°C in 1 month whilethat at pH 9 remained above 95%, apparently dueto the difference in the rate of sucrose hydrolysisat different pHs (te Booy et al., 1992; Skrabanjaet al., 1994).

5.2.4. Moisture contentThe residual moisture content after lyophiliza-

tion often controls long-term protein stability,both physically and chemically (Franks, 1990;Hatley and Franks, 1991). The moisture contentof a lyophilized protein formulation may changesignificantly during storage due to a variety offactors, such as stopper moisture release and leak-age (Section 5.4), crystallization of an amorphousexcipient (next section), or moisture release froman excipient hydrate (Yu et al., 1999). Water canaffect protein stability both indirectly as a plasti-cizer or reaction medium and directly as a reac-tant or a product (Shalaev and Zografi, 1996).

Indirectly as a plasticizer, water drastically de-creases glass transition temperature of proteins,polymers or other formulation excipients (Slade etal., 1989; Roos and Karel, 1991; Buera et al.,1992; te Booy et al., 1992; Roos, 1993; Wolkers etal., 1998a). The quantitative effect of water onglass transition temperature can be easily esti-mated by the Gordon–Taylor equation (Hancock

and Zografi, 1994). Depression of Tg by watermay reach 10° or more for each percent of mois-ture retained, especially at low-level moisture con-tents (Angell, 1995; Hatley, 1997; Rossi et al.,1997). Therefore, a lyophilized protein may easilyadsorb sufficient amounts of moisture during stor-age to reduce its Tg below the storage tempera-ture, to accelerate its instability, and to causepossible product collapse (Oksanen and Zografi,1990). For example, dimerization of lyophilizedinsulin increases significantly when water contentis high enough to cause significant decrease inglass transition temperature and cake collapse(Strickley and Anderson, 1997). High moisturecontent also facilitates crystallization of formula-tion excipients such as various sugars, which willbe discussed in the next section.

Generally, increasing moisture content of alyophilized protein increases the deterioration rateof proteins. Sorbed water increases the free vol-ume of a lyophilized protein, promoting molecu-lar mobility (Towns, 1995; Shamblin et al., 1998).The denaturation temperature (161°C) of driedsomatotropin decreased with increasing moisturecontent until it reached a plateau of 65°C atmoisture contents of \28% (Bell et al., 1995).Increasing moisture content of lyophilized tPAfrom 4.6 to 7.6 or 18.0% increased loss of proteinactivity during storage at 50°C (Hsu et al., 1991).Increasing the moisture content of a lyophilizedinvertase formulation from 2.1 to 10.9% graduallydecreased the stability of the protein upon incuba-tion at 90°C for 10 h (Cardona et al., 1997).Increasing relative humidity accelerated activityloss of vacuum-dried restriction enzyme EcoRI inboth sucrose and trehalose formulations duringstorage at 45°C (Rossi et al., 1997). Protein aggre-gation often increases with increasing relative hu-midity during storage, such as insulin (Costantinoet al., 1994b) and lysozyme (Separovic et al.,1998). The aggregation-induced loss of activity ofbovine pancreas RNase was faster at 9.8% mois-ture than at 1.9% (Townsend and DeLuca, 1990).Moisture uptake was also the cause of reduceda-helical conformation and increased b-structure(a common theme of protein aggregation) ofspray-dried rhG–CSF or recombinant consensusinterferon (rConIFN) (French et al., 1995).


The increased deterioration in lyophilizedproteins at high moisture contents may resultfrom increased chemical degradations. The forma-tion of degradation products of insulin inlyophilized formulations was directly proportionalto the moisture content of the formulation from 3to 52% (Strickley and Anderson, 1996, 1997). Theincreased rate of a chemical reaction at high watercontents is due to increased mobility of waterinvolved in the reaction and its positive effect onthe mobility of proteins (Hageman, 1988; Towns,1995). The diffusion coefficient (mobility) of waterin PVP at 25°C can increase exponentially withincreased motion of PVP side chains as the watercontent increases (Oksanen and Zografi, 1993). Infact, the rate of hydrogen exchange in lysozymepowder containing 0.2 g water g−1 protein (about17% moisture) is the same as that in a dilutesolution (Rupley and Careri, 1991).

In many cases, however, the effect of moistureon protein stability is a complex function. Changet al. (1996c) studied the stability of lyophilizedrhIL-1ra (from 50 mg ml−1 rhIL-1ra, 2% Glycine,and 1% sucrose in 10 mM citrate buffer at pH6.5) at moisture levels between 0.5 and 3.2%(w/w) at 30°C. They found that the least stableformulation had a moisture level of around 0.8%and the more stable formulations had a moisturelevel of either 50.5 or 3.2%. In a few studies, aclear bell-shaped relationship was demonstratedbetween protein stability and moisture content.For example, maximum aggregation (78%) oflyophilized tetanus toxoid occurred at a watercontent of about 36% during storage at 37°C for10 days, and less aggregation was observed atwater contents either below or above that level(Schwendeman et al., 1995). Aggregation of both(lyophilized) rHA and BSA during storage at37°C had a bell-shaped relationship as a functionof water content with maximum aggregation atabout 32 and 28% moisture, respectively (Liu etal., 1990; Costantino et al., 1995b). Similar bell-shaped aggregation dependence on water contentwas also observed for ovalbumin (chicken eggalbumin), glucose oxidase, bovine b-lactoglobulin(Liu et al., 1990), and insulin (Katakam andBanga, 1995; Separovic et al., 1998).

Changing moisture content in a protein formu-lation may change degradation mechanisms andthus, the overall rate of degradation. The mecha-nism of aggregation of lyophilized tetanus toxoid(150 kD) was different under different relativehumidities during storage. Under 80% RH, all theaggregates were formed by non-disulfide covalentbonding after storing the lyophilized protein at37°C for 10 days, but under 97% RH, 55% of theaggregates were formed by hydrophobic interac-tion and scrambling of disulfide bonds (Schwen-deman et al., 1996). Hekman et al. (1995) foundthat increasing moisture content of lyophilizeddiethylenetriaminepentaacetic anhydride (DTPA)-conjugated IgG above 20% could change the ma-jor degradation pathway from Maillard reactionto one which also involved precipitation duringstorage.

The effect of residual moisture on protein sta-bility can be strongly influenced by several otherfactors, including temperature, excipient composi-tion, and moisture distribution within a productcake. Pikal et al. (1992) demonstrated thatlyophilized rGH, in the presence of glycine andmannitol, aggregated almost linearly with increas-ing moisture content from 0.7 to 2.5% duringstorage at 40°C, but the curve was bell-shaped at25°C. The rate of hGH degradation in thelyophilized formulation varied greatly with for-mulation composition as well as with storagetemperature. They also found that hGH degrada-tion in a lyophilized formulation containingglycine and mannitol was more sensitive to highlevels of moisture or headspace oxygen than in aformulation without these excipients (Pikal et al.,1992). Moisture distribution within a proteinproduct is usually quite uneven and this mayresult in a biphasic loss of activity (initial fastphase and remaining slow phase) during storage(Franks, 1990).

5.2.5. Type and concentration of formulationexcipients

Many formulation excipients stabilize proteinsin solid state in a concentration-dependent man-ner (see Section 5.3). On the other hand, overuseof an excipient(s) may eventually destabilize aprotein. Even sugars/polyols, the universal protein


stabilizers, may destabilize a protein if their quan-tities are not appropriately used in a proteinformulation (see next section and Section 5.3).

Certain excipients, that are used to stabilizeproteins in liquid state or during lyophilization,may destabilize proteins in solid state. Ascorbicacid is a frequently used antioxidant for smalldrugs, but at 5 mM, it reduced the storage stabil-ity of lyophilized elastase (20 mg ml−1 in 10 mMsodium acetate, pH 5.0) drastically at 40°C and79% RH (Chang et al., 1993).

Lyophilized protein formulations often containa buffering agent(s). Different proteins may needdifferent buffering agents for maximum stabiliza-tion in solid state. For example, histidine has beenshown to be the best buffer agent for minimizingaggregation of lyophilized recombinant factor IX(rFIX) during storage at 30°C among all thebuffering agents examined, including sodiumphosphate, potassium phosphate, and Tris (Bushet al., 1998). Sodium phosphate has been shownto be a better buffering agent than sodium citrateand sodium maleate in maintaining the solubilityof lyophilized rIFN-b-1b (Betaseron®) upon re-constitution (Lin et al., 1996). In a different study,glycocholate buffer was found to be better thansuccinate buffer in stabilizing lyophilized IFN-gas about 50% IFN-g activity was lost in alyophilized succinate-containing formulation dur-ing storage at 25°C for 4 weeks, while about 28%activity was lost in glycocholate-containing for-mulation (Lam et al., 1996).

The buffer concentration also influences storagestability of lyophilized proteins. The loss of activ-ity of lyophilized bovine pancreatic RNase insodium phosphate buffer (pH 6.4) during storageat 45°C increased with increasing buffer concen-tration in the range of 0, 0.02, and 0.2 M(Townsend and DeLuca, 1990). One possiblecause is that buffering agents may change Tg oflyophilized formulations. The Tg of lyophilizedrhIL-1ra formulation containing 1% sucrose, 4%mannitol, and 2% glycine decreased from 46 to26°C when the buffering agent sodium citrate wasreplaced with sodium phosphate (Chang et al.,1996a). A lower Tg generally leads to a less stableformulation (see Section 5.2).

5.2.6. Crystallization of amorphous excipientsThe potential for crystallization of amorphous

excipients in a lyophilized protein formulationduring storage always exists because the crys-talline state is more stable thermodynamically(Hancock and Zografi, 1997). Crystallization usu-ally destabilizes a protein due to a loss of intimateexcipient interaction with proteins and a possibledecrease in Tg of the amorphous phase. The de-crease in Tg arises from increased moisture con-tent of the amorphous phase because the relativeamount of the amorphous material(s) is reducedafter crystallization and the total amount of mois-ture in a sealed container does not change. How-ever, if the amorphous excipient crystallizes ashydrates, such as trehalose dihydrate or raffinosepentahydrates, the Tg of the amorphous phase canactually increase (Aldous et al., 1995).

There are many cases of protein destabilizationdue to excipient crystallization. Inositol crystal-lization in a lyophilized b-galactosidase formula-tion destabilized the protein during storage(Izutsu et al., 1994a). The formation of bFGFdegradants in a lyophilized formulation under‘acidic’ condition was accelerated by sucrose crys-tallization (Wu et al., 1998). Mannitol crystalliza-tion was apparently the cause of a drasticdecrease in storage stability for co-spray-driedanti-IgE monoclonal antibody at 5 or 30°C(Costantino et al., 1998a). Crystallization of ex-cipients also causes possible ‘pH’ shifting, affect-ing protein stability as observed in lyophilizedbFGF during storage (Wu et al., 1998).

Many sugar/polyol excipients have a tendencyto crystallize during storage. The rate of crystal-lization increases with increasing temperature orrelative humidity (Schmitt et al., 1999). Immediatecrystallization of disaccharides can occur whenthese sugars are heated to about 50°C above theirTgs (Cardona et al., 1997). Lyophilized sucrose ina LDH formulation (in 10 mM PBS, pH 7.2) ismainly amorphous but tends to crystallize withincreasing residual water content or during heat-ing or storage at elevated temperatures (Moreiraet al., 1998). Spray-dried amorphous lactose, su-crose, or trehalose can easily crystallize at 25°Cand under RH of ]52% (Naini et al., 1998).Both lyophilized and spray-dried rhDNase–lac-


tose mixture could easily crystallize upon expo-sure to high humidity (]60%) (Chan and Gonda,1998; Chan et al., 1999). An amorphous sucroseformulation could change to a crystalline formcompletely in 1 month at 60°C (te Booy et al.,1992). Because of sucrose crystallization, trehalosestabilized lyophilized rFXIII and Humicola lanug-inosa lipase much more effectively than sucroseduring storage at 60°C (Kreilgaard et al., 1998a,1999). In a more detailed study, it was demon-strated that increasing moisture content droppedroughly linearly the crystallization temperature ofsucrose or recombinant human somatotropin(rbSt) (Sarciaux and Hageman, 1997). Therefore,moisture content in a protein formulation shouldbe minimized to prevent crystallization of anexcipient(s).

Crystallization tendency of amorphous excipi-ents is strongly affected by their relative amountin a protein formulation. For example, amor-phous mannitol, sorbitol, or methyl a-D-mannopyranoside in a rhGH formulation at anexcipient:rhGH molar ratio of less than 131:1 didnot crystallize easily during storage at 50°C for 4weeks, but crystallization occurred at 1000:1 mo-lar ratio (Costantino et al., 1998b). Increasing therelative content of sucrose significantly decreasedits crystallization temperature in a lyophilized su-crose:rhGH formulation (Costantino et al., 1998c)and facilitated its crystallization under moisture ina spray-dried sucrose:trypsinogen formulation(Tzannis and Prestrelski, 1999b). Mannitol at 10or 20% stabilized a co-spray-dried anti-IgE mono-clonal antibody during storage at 5 or 30°C, butmannitol at 30% crystallized and drastically desta-bilized the protein during storage (Costantino etal., 1998a). Similarly, inositol at concentrationsbetween 50 and 160 mM in a lyophilized b-galac-tosidase formulation did not crystallize duringstorage at 70°C for 7 days but crystallized atconcentrations above 250 mM (Izutsu et al.,1994a). DSC data indicated that the crystalliza-tion temperature of inositol decreased from 110 to60°C when the inositol concentration in the for-mulation increased from 50 to 500 mM.

Crystallization tendency of amorphous excipi-ents also changes depending on the type andconcentration of co-existing formulation excipi-

ents (also see Section 5.4). Inclusion of sodiumphosphate in a spray-dried formulation contain-ing mannitol and an anti-IgE monoclonal anti-body inhibited mannitol crystallization, which inturn reduced solid-state protein aggregation dur-ing storage at 5 or 30°C (Costantino et al.,1998a). Addition of phosphate and citrate in alyophilized sucrose formulation inhibited the crys-tallization of sucrose as demonstrated by DSC (teBooy et al., 1992).

Many polymer excipients (or proteins) havebeen shown to inhibit the crystallization of smallcarbohydrate excipients. Maltodextrins or PVPeffectively retard the crystallization of sucrose(Roos and Karel, 1991; Shamblin et al., 1996).Dextran or carboxymethyl cellulose sodium salt(CMC-Na) effectively inhibits the crystallizationof inositol (Izutsu et al., 1994a). PVP can reducethe tendency of moisture-induced crystallizationof lyophilized sucrose (Shamblin and Zografi,1999). Both catalase and insulin can inhibit lac-tose crystallization in a co-spray-dried amorphousmixture upon exposure to short-term elevated hu-midity (Forbes et al., 1998).

To examine excipient crystallization in aprotein formulation during storage, X-ray diffrac-tometry or IR can be used (Izutsu et al., 1994a;Kreilgaard et al., 1999). Recently, Raman spec-troscopy has been shown to detect 1% amorphousor crystalline content and proved to be anotheruseful method for crystallinity detection (Taylorand Zografi, 1998a).

5.2.7. Reconstitution mediumThe reconstitution step may potentially affect

protein stability by several mechanisms. First,rapid reconstitution with water may not allow adried protein to rehydrate as slowly as the dehy-dration step (Cleland et al., 1993). Therefore, aprotein may not be able to refold to its nativeform during reconstitution, causing denaturationand/or aggregation. Second, the pH of water-re-constituted protein formulations may be differentfrom those before lyophilization, possibly due toloss of some formulation components. For exam-ple, three insulin solutions have pHs of 2.0, 3.0and 4.0 before lyophilization, and the respectivepHs of the reconstituted lyophilized insulin were


found to be 3.1, 3.3 and 4.1, presumably due toloss of HCl during lyophilization (Strickley andAnderson, 1996). Lyophilization-induced evapo-ration of acetic acid can also cause similar pHchanges (Hatley et al., 1996). Last, many proteinsare temperature-sensitive and the temperature ofthe reconstitution medium can make a significantdifference. For instance, the activity of H+-AT-Pase was completely recovered when it was rehy-drated at 20°C but the recovered activity wasreduced to 62% when it was rehydrated at 30°C(Sampedro et al., 1998).

To alleviate these problems, addition of a sur-factant and/or a stabilizer(s) in the reconstitutionmedium or adjustment of medium pH may berequired. For example, inclusion of Tween 20 inthe reconstitution medium decreased aggregationof interferon-g compared with reconstitution withwater alone (Webb et al., 1998). Addition ofmaltose in the reconstitution buffer increased ac-tivity recovery of the dried restriction endonucle-ase HindIII (Uritani et al., 1995). The amount ofaggregates in reconstituted IL-2 formulationcould be reduced by inclusion of one of thefollowing excipients in the medium: heparin, dex-tran sulfate, glycine, lysine–HCl, polylysine,Tween 20, or HP-b-CD (Zhang et al., 1996).Lowering reconstitution solution pH from 7.0 to4.0 also lowered the amount of aggregates afterreconstitution. About 11% of recombinant humankeratinocyte growth factor (rhKGF) formed ag-gregates immediately after reconstitution of thelyophilized KGF with water, but only about 1.5%formed aggregates after reconstitution with watercontaining 0.05% (w/v) heparin (16 kD) or su-crose octasulfate (SOS) (Zhang et al., 1995).Other stabilizers that inhibited reconstitution-in-duced aggregation included dextran sulfate, fu-coidan, pentosan polysulfate, chondroitin sulfate,myo-inositol sulfate, sulfated b-cyclodextrin,polyphosphoric acid, NaCl, (NH)2SO4, etc.

5.3. Stabilization of solid proteins by excipients

Solid protein pharmaceuticals may need to bestabilized in two stages, during lyophilization(preparation); and during long-term storage.Protein stabilization during lyophilization has

been discussed in Section 3. The following sectionfocuses on protein stabilization for long-termstorage.

5.3.1. Stabilization mechanismsStabilization mechanisms for lyophilized

proteins during long-term storage are similar tothose for protein lyoprotection (see Section 3.2).These stabilization mechanisms include formationof an amorphous glassy state, water replacementby excipients, and hydrogen bonding between ex-cipients and proteins (Fox, 1995). It has beenshown that excipients capable of replacing watermolecules upon dehydration better preserve thenative structure of proteins, resulting in enhancedstability (Prestrelski et al., 1995). Nevertheless, itis now clear that glass formation alone is notsufficient for protein stabilization in solid state.The supporting evidence for this argument is thatglass-forming polymers fail to stabilize or evendestabilize solid proteins during storage (Colacoet al., 1992; Nakai et al., 1998; Lueckel et al.,1998b). Therefore, a combination of these mecha-nisms is required for maximum protein stabiliza-tion in solid state (Crowe et al., 1996, 1998).

In addition to the above-mentioned mecha-nisms, other stabilization hypotheses have alsobeen proposed. One of them is the prevention ofprotein–protein interaction and aggregation byphysical dilution and separation of proteinmolecules. The reduced aggregation of lyophilizedBSA in the presence of NaCl, sodium phosphate,carboxymethyl-cellulose, dextran, DEAE-dextran,or PEG on incubation at 37°C has been attributedto a general dilution effect (Liu et al., 1990). Thismechanism has also been suggested in stabiliza-tion of other proteins during storage such as rHAby dextran (Costantino et al., 1995a), IL-1ra bysucrose (Chang et al., 1996c), and insulin bytrehalose (inhibition of covalent dimerization) at35°C (Strickley and Anderson, 1997).

Another stabilization hypothesis is excipient–water interactions. Costantino et al. (1995a)demonstrated that rHA co-lyophilized with NaCl(NaCl:protein=1:6 on a weight basis) did notexhibit any aggregation after a 4-day incubationperiod at 37°C and 96% RH, while the proteinwithout the excipient lost over 80% solubility after


just 1 day under the same conditions. Since inclu-sion of NaCl did not induce any significantchange in the secondary structure of rHA afterlyophilization, the stabilization effect of NaCl wasapparently due to water uptake by NaCl in thevicinity of rHA, which facilitated protein refold-ing into its native and more stable conformation.This stabilization mechanism was clearly demon-strated in a similar study, in which it was foundthat the effectiveness of several excipients (D-glu-caric acid, D-gluconic acid, sorbitol, sodium chlo-ride, etc.) in inhibiting aggregation of lyophilizedrHA roughly correlated with their water uptakeunder 96% RH (Costantino et al., 1995b). Thegreater the excipient’s affinity for water, generally,the greater the stabilizing effect. However, if aprotein is very sensitive to moisture, this stabiliza-tion mechanism is not applicable. For example,the aggregation of moisture-sensitive TT in alyophilized formulation containing NaCl, D-sor-bitol, or PEG 20 000 (at an excipient/proteinweight ratio of 1:5) did not correlate with wateruptake in these formulations during storage at37°C and 86% RH (Schwendeman et al., 1995).

Polymers may stabilize proteins by increasingTg of a solid protein formulation, because poly-mers usually have higher Tgs due to their highmolecular weights (te Booy et al., 1992; Prestrelskiet al., 1995). The Tgs of maltodextrin and PVPwere found to increase linearly with their molecu-lar weights, as described by the Fox and Floryequation (Roos and Karel, 1991; Buera et al.,1992). Because of this effect, the Tg of dextran-formulated g-globulin formulations increased sig-nificantly with increasing molecular weight ofdextran from 10 to 510 kD (Yoshioka et al.,1997). In addition, polymers (or proteins) mayindirectly stabilize proteins by inhibiting the crys-tallization of other stabilizing excipients in a solidformulation (see Section 5.2).

5.3.2. Sugars/polyolsSugars or polyols have commonly been used to

stabilize lyophilized proteins for long-term stor-age, such as sucrose, trehalose, mannitol, sorbitol,etc. Sucrose has been shown to prevent aFGFaggregation completely at 2% during storage at25°C for 1 year (Volkin and Middaugh, 1996), to

inhibit the aggregation rate of rhIL-1ra signifi-cantly in a concentration range of 0–10% (w/v)(Chang et al., 1996a,c), and to stabilize rFXIII at100 mM during storage at 40°C (Kreilgaard et al.,1998a). Sucrose has also been used to improve thestorage stability of factor IX (FIX) (Bush et al.,1998) and rIFN-b-1b (Betaseron®) (Lin et al.,1996). Trehalose increased the stability oflyophilized invertase during incubation at 90°C(Schebor et al., 1996; Cardona et al., 1997). Bothsucrose and trehalose inhibited the dimer forma-tion in TNF during storage at 37°C (Hora et al.,1992a). Trehalose or lactose at or above 300:1(excipient:protein) molar ratio inhibited aggrega-tion of a spray-dried anti-IgE monoclonal anti-body effectively during storage at 30°C underboth 11 and 38% RH (Andya et al., 1999). Man-nitol has been shown to decrease aggregation oflyophilized TT during storage at 37°C and 86%RH (Costantino et al., 1996) and lyophilized anti-body-vinca conjugate during storage at 25°C (Royet al., 1992). Sorbitol could prevent aggregationof rHA completely at 1:1 (excipient:protein)weight ratio during storage at 37°C for 4 days(Costantino et al., 1995b) and reduce aggregationof lyophilized TT effectively at 1:5 weight ratioduring storage at 37°C and 86% RH for 6 days(Schwendeman et al., 1995). Inositol was able toprotect lyophilized b-galactosidase in a concentra-tion-dependent manner during storage at temper-atures below 50°C (Izutsu et al., 1994a). A varietyof mono-, di-, and trisaccharides or polyols pro-tected lyophilized restriction enzyme PstI to vari-ous degrees during storage at 37°C (Colaco et al.,1992). More examples can be found in Table 2.

As used in lyoprotection, sucrose and trehaloseseem to be the most commonly used disaccharidesfor protection of solid proteins during long-termstorage. Their relative effect in stabilizing solidproteins is still a much-debated subject. Clearly,trehalose has several advantages over sucrose, asdiscussed in Section 3.1. In addition, formation ofa dihydrate during storage of freeze-dried tre-halose has been demonstrated to prevent mois-ture-induced decrease in glass transitiontemperature, which usually destabilizes proteins(Aldous et al., 1995; Crowe et al., 1996). Theseadvantages might explain why trehalose was


Tab

le2

Exa

mpl

esof

prot

ein

inst

abili

tyan

dst

abili

zati

onby

exci

pien

ts

Pro

tein

sR

efer

ence

sF

orm

ulat

ion

com

posi

tion

sSt

udy

cond

itio

nsR

emai

ning

Act

ivit

y(o

ras

stat

ed)

�24

%10

mg

ml−

1ac

tin

in1

mM

Tri

s(p

H8.

0),

Alli

son

etal

.,19

98F

reez

e-dr

ying

Act

in0.

1m

MA

TP

,0.

25m

M2

mer

capt

oeth

anol

,0.

1m

MC

aCl 2

,an

d0.

0025

%N

aN3

33%

+1%

(w/v

)su

cros

e+

5%(w

/v)

sucr

ose

�67

%55

%R

amos

etal

.,19

97A

DH

(bak

er’s

yeas

t)50

mgm

l−1

in20

mM

pota

ssiu

mph

osph

ate

Fre

eze-

dryi

ngbu

ffer

(pH

7.6)

+50

0m

MK

Cl

34%

98%

+50

0m

Mtr

ehal

ose

+50

0m

Mm

anno

sylg

lyce

rate

89%

Bili

rubi

nox

idas

e0.

4m

gm

l−1

in0.

2m

gm

l−1

Tri

ton

X-1

00F

reez

e-dr

ying

�56

%N

akai

etal

.,19

98an

d40

mg

ml−

1de

xtra

n�

58%

0.4

mg

ml−

1in

0.2

mg

ml−

1T

rito

nX

-100

and

40m

gm

l−1

PV

A�

85%

0.4

mg

ml−

1in

0.2

mg

ml−

1T

rito

nX

-100

and

40m

gm

l−1

treh

alos

eSh

ikam

aan

dY

amaz

aki,

8.4

mgm

l−1

in10

mM

phos

phat

ebu

ffer

80%

Fre

ezin

gfo

r10%

atte

mpe

ratu

res

Cat

alas

e(o

xliv

er)

1961

(pH

7.0)

betw

een

−15

and

−70

°C95

%+

1M

glyc

erol

90%

+0.

5m

Mso

dium

dode

cyl

sulf

ate

1m

gm

l−1

in50

mM

phos

phat

ebu

ffer

Fre

eze-

dryi

ng�

35%

Tan

aka

etal

.,19

91C

atal

ase

(bee

fliv

er)

+0.

2m

gm

l−1

mal

tose

�70

%+

1m

gm

l−1

mal

tose

�90

%�

90%

+1.

3m

g−

1m

lm

alto

seE

last

ase

Incu

bati

onat

40°C

,79

%R

Hfo

r2

Lyo

phili

zed

form

ulat

ion

at20

mg

ml−

1in

1033

%C

hang

etal

.,19

93w

eeks

mM

sodi

umac

etat

e(p

H5.

0)+

10%

sucr

ose

86%

82%

.+

10%

dext

ran

40Iz

utsu

etal

.,19

94a

20mg

ml−

1in

200

mM

phos

phat

ebu

ffer

(7.4

)F

reez

e-dr

ying

47%

b-G

alac

tosi

dase

and

100

mM

inos

itol

+40

0m

Min

osit

ol�

90%

�10

0%+

400

mM

inos

itol

and

1m

gm

l−1

dext

ran

+40

0m

Min

osit

olan

d1

mg

ml−

1C

MC

-Na

�10

0%Iz

utsu

etal

.,19

93b-

Gal

acto

sida

seF

reez

e-dr

ying

2mg

ml−

1in

10m

Mso

dium

phos

phat

e,(p

H7.

4)�

5%+

50m

Mm

anni

tol

�95

%�

70%

+20

0m

Mm

anni

tol

+50

0m

Mm

anni

tol

�25

%


Tab

le2

(Con

tinu

ed)

Pro

tein

sR

efer

ence

sF

orm

ulat

ion

com

posi

tion

sSt

udy

cond

itio

nsR

emai

ning

Act

ivit

y(o

ras

stat

ed)

�5%

Fre

eze-

dryi

ngb-

Gal

acto

sida

seIz

utsu

etal

.,19

932

mgm

l−1

in20

0m

Mso

dium

phos

phat

e,pH

7.4

+50

mM

man

nito

l�

40%

�95

%+

200

mM

man

nito

l+

500

mM

man

nito

l�

95%

82%

rFX

III

Kre

ilgaa

rdet

al.,

2m

gm

l−1

in0.

1m

ME

DT

Aan

d10

mM

Tri

sF

reez

e-dr

ying

buff

er(p

H8)

1998

a90

%+

100

mM

man

nito

l+

1%(w

/v)

PE

G11

5%10

9%+

3.5%

(w/v

)de

xtra

n+

100

mM

treh

alos

e11

7%+

100

mM

sucr

ose

111%

106%

+0.

002%

(w/v

)T

wee

n20

Ram

oset

al.,

1997

50mg

ml−

1in

20m

Mpo

tass

ium

phos

phat

eF

reez

e-dr

ying

65%

GD

Hbu

ffer

(pH

7.5)

32%

+50

0m

MK

Cl

+50

0m

Mtr

ehal

ose

86%

+50

0m

Mm

anno

sylg

lyce

rate

95%

Lyo

phili

zed

form

ulat

ion

at2

mg

ml−

1In

cuba

tion

at50

°CR

ate

cons

tant

=4.

5pe

rda

yhG

HC

osta

ntin

oet

al.,

1998

b0.

004

per

day

+m

anni

tol

at31

:1m

olar

rati

o+

treh

alos

eat

31:1

mol

arra

tio

0.00

2pe

rda

y+

cello

bios

eat

31:1

mol

arra

tio

0.00

03pe

rda

ySa

mpe

dro

etal

.,H

+-A

TP

ase

Fre

eze-

dryi

ng1

mg

ml−

1in

1m

ME

GT

A-T

ris

buff

er(p

H7.

0)4%

1998

100%

+20

mg

treh

alos

epe

rm

gpr

otei

n+

20m

gm

alto

sepe

rm

gpr

otei

n91

%+

20m

gsu

cros

epe

rm

gpr

otei

n84

%+

20m

ggl

ucos

epe

rm

gpr

otei

n72

%+

20m

gga

lact

ose

per

mg

prot

ein

37%

Lyo

phili

zed

form

ulat

ion

at10

mg

ml−

1rh

IL-1

ra,

Incu

bati

onat

50°C

rhIL

-1ra

Cha

nget

al.,

1996

cD

egra

dati

onat

1.8%

per

2%gl

ycin

ean

d10

mM

sodi

umci

trat

e(p

H6.

5)w

eek,

aggr

egat

ion

at0.

08(A

500n

m)

per

wee

k0.

15%

and

0.00

8,re

spec

tive

ly+

1%su

cros

e(w

/v)

6.4%

and

0.00

3,re

spec

tive

ly+

1%m

alto

se+

1%so

rbit

ol2.

3%an

d0.

04,

resp

ecti

vely

+1%

treh

alos

e0.

43%

and

0.00

5,re

spec

tive

ly47

%P

rest

rels

kiet

al.,

Incu

bati

onat

45°C

for

4w

eeks

IL-2

Lyo

phili

zed

form

ulat

ion

at0.

8m

gm

l−1

inM

ES

1995

buff

er(p

H7)

+0.

5%su

cros

e86

%+

5%m

anni

tol

28%


Tab

le2

(Con

tinu

ed)

Rem

aini

ngA

ctiv

ity

Ref

eren

ces

Stud

yco

ndit

ions

Pro

tein

sF

orm

ulat

ion

com

posi

tion

s(o

ras

stat

ed)

Lyo

phili

zed

form

ulat

ion

at1

mg

ml−

1(p

H4)

Stri

ckle

yan

d0.

04m

gm

l−1

degr

adan

tIn

sulin

Incu

bati

onat

35°C

for

7da

ysA

nder

son,

1997

(18%

dim

ers)

+5

mg

ml−

1tr

ehal

ose

0.02

mg

ml−

1de

grad

ant

(1%

dim

ers)

Anc

hord

oquy

and

76%

Fre

eze-

thaw

ing

LD

HA

bout

0.5

mg

ml−

1(0

.25

mg

ml−

1ra

bbit

mus

cle

Car

pent

er,

1996

and

0.25

mg

ml−

1po

rcin

ehe

art

LD

H)

in0.

1M

NaC

lan

d10

mM

pota

ssiu

mph

osph

ate,

pH7.

5)80

%+

0.1%

(w/v

)B

SA+

1.0%

(w/v

)B

SA10

0%+

0.1%

(w/v

)P

VP

79%

+1.

0%(w

/v)

PV

P98

%L

DH

(rab

bit

mus

cle)

Ald

enan

dF

reez

e-th

awin

g19

%25

mgm

l−1

inw

ater

Mag

nuss

on,

1997

+1%

PE

G79

%9

mgm

l−1

inw

ater

Fre

eze-

thaw

ing

36%

LD

H(r

abbi

tm

uscl

e)N

ema

and

Avi

s,19

9282

%+

1%(w

/v)

met

hoce

l+

1%(w

/v)

Plu

roni

cF

127

87%

76%

+2.

5%(w

/v)

PV

P+

0.2

MP

EG

400

85%

+0.

5%(w

/v)

gela

tin

91%

100%

+1%

(w/v

)B

SA+

0.9%

(w/v

)b-

cycl

odex

trin

31%

+5%

(w/v

)de

xtra

n76

%+

5%(w

/v)

treh

alos

e26

% 6%+

5%(w

/v)

man

nito

l+

5%(w

/v)

sucr

ose

73%

+10

%(w

/v)

sucr

ose

88%

+34

.2%

(w/v

,1

M)

sucr

ose

102%

+0.

05%

(w/v

)B

rij

3010

0%+

0.00

2%T

wee

n80

57%

65%

+0.

005%

Tw

een

80+

0.05

%T

wee

n80

79%

+0.

1%T

wee

n80

81%

80%

+0.

5%T

wee

n80

+1%

Tw

een

8083

%2

mgm

l−1

in50

mM

sodi

umph

osph

ate

�9%

Izut

suet

al.,

1995

Fre

eze-

dryi

ngL

DH

buff

er(p

H7.

4)


Tab

le2

(Con

tinu

ed)

Pro

tein

sR

efer

ence

sF

orm

ulat

ion

com

posi

tion

sSt

udy

cond

itio

nsR

emai

ning

Act

ivit

y(o

ras

stat

ed)

�14

%+

400

mM

man

nito

l+

200

mM

sucr

ose

�30

%+

10m

gm

l−1

CH

AP

S�

76%

+10

mg

ml−

1H

P-b

-CD

�73

%+

10m

gm

l−1

sodi

umch

olat

e�

60%

+10

mg

ml−

1P

EG

3000

�30

%�

38%

+10

mg

ml−

1P

EG

2000

0F

reez

e-dr

ying

66%

Anc

hord

oquy

and

Abo

ut0.

5m

gm

l−1

(0.2

5m

gm

l−1

rabb

itm

uscl

eL

DH

and

0.25

mg

ml−

1po

rcin

ehe

art

LD

H)

in0.

1M

Car

pent

er,

1996

NaC

lan

d10

mM

pota

ssiu

mph

osph

ate,

(pH

7.5)

65%

+0.

1%(w

/v)

BSA

86%

+1.

0%(w

/v)

BSA

92%

+10

.0%

(w/v

)B

SA77

%+

0.1%

(w/v

)P

VP

+1.

0%(w

/v)

PV

P70

%+

1.0%

(w/v

)P

VP

97%

50mg

ml−

1in

50m

MT

ris–

HC

l(p

H7.

2)F

reez

e-dr

ying

LD

H27

%A

nder

sson

and

Hat

ti-K

aul,

1999

+0.

01%

poly

ethy

lene

imin

e(l

owM

W)

43%

+0.

1%po

lyet

hyle

neim

ine

(low

MW

)47

%69

%+

1%po

lyet

hyle

neim

ine

(low

MW

)+

0.01

%po

lyet

hyle

neim

ine

(hig

hM

W)

52%

+0.

1%po

lyet

hyle

neim

ine

(hig

hM

W)

64%

69%

+1%

poly

ethy

lene

imin

e(h

igh

M.W

.)L

yoph

ilize

dfo

rmul

atio

nat

50mg

ml−

1in

6%A

nder

sson

and

Incu

bati

onat

36°C

for

1w

eek

LD

HH

atti

-Kau

l,19

9950

mM

Tri

s–H

Cl

(pH

7.2)

+0.

01%

poly

ethy

lene

imin

e(l

owM

W)

27%

+0.

1%po

lyet

hyle

neim

ine

(low

MW

)31

%+

1%po

lyet

hyle

neim

ine

(low

MW

)59

%+

0.01

%po

lyet

hyle

neim

ine

(hig

hM

W)

21%

51%

+0.

1%po

lyet

hyle

neim

ine

(hig

hM

W)

+1%

poly

ethy

lene

imin

e(h

igh

MW

)64

%L

DH

Pre

stre

lski

etal

.,2

mg

ml−

1in

20m

Mpo

tass

ium

phos

phat

eF

reez

e-dr

ying

46%

1993

a(r

abbi

tm

uscl

e)bu

ffer

(pH

7.5)

46%

+1%

PE

G39

%+

10m

Mm

anni

tol

+10

mM

lact

ose

43%

+10

mM

treh

alos

e39

%+

1%P

EG

and

10m

Mm

anni

tol

81%


Tab

le2

(Con

tinu

ed)

Pro

tein

sR

efer

ence

sF

orm

ulat

ion

com

posi

tion

sSt

udy

cond

itio

nsR

emai

ning

Act

ivit

y(o

ras

stat

ed)

+1%

PE

Gan

d10

mM

lact

ose

100% 90

%+

1%P

EG

and

10m

Mtr

ehal

ose

Ram

oset

al.,

1997

50mg

ml−

1in

20m

Mpo

tass

ium

phos

phat

ebu

ffer

Fre

eze-

dryi

ng12

%L

DH

(rab

bit

mus

cle)

(pH

7.6)

0%+

500

mM

KC

l+

500

mM

treh

alos

e54

%+

500

mM

man

nosy

lgly

cera

te85

% 0%P

rest

rels

ki,

etal

.,F

reez

e-dr

ying

PF

K2

mg

ml−

1in

20m

Mpo

tass

ium

phos

phat

e19

93a

buff

er(p

H8.

0)20

%+

1%P

EG

+10

mM

man

nito

l5%

+10

mM

lact

ose

5% 0%+

10m

Mtr

ehal

ose

+1%

PE

Gan

d10

mM

man

nito

l30

%10

3%+

1%P

EG

and

10m

Mla

ctos

e+

1%P

EG

and

10m

Mtr

ehal

ose

100%

PF

KF

reez

e-th

awin

g50

mgm

l−1

in10

mM

pota

ssiu

mph

osph

ate

Car

pent

eran

d0%

Cro

we,

1989

(pH

8.0)

and

5m

Mdi

thio

thre

itol

+15

0m

gm

l−1

treh

alos

e�

65%

+30

0m

gm

l−1

treh

alos

e�

90%

PF

K25

mgm

l−1

in1

mM

sodi

umbo

rate

(pH

7.8)

,F

reez

e-dr

ying

0%C

arpe

nter

etal

.,19

8625

mM

K2SO

4,

2m

M(N

H4) 2

SO4,

and

5m

Mdi

thio

thre

itol

+1

mM

ZnS

O4

B13

%B

10%

+50

mM

treh

alos

e,su

cros

e,or

mal

tose

+1

mM

ZnS

O4

and

50m

Mtr

ehal

ose

�80

%+

1m

MZ

nSO

4an

d50

mM

sucr

ose,

orm

alto

se�

60%

+1

mM

ZnS

O4

and

50m

Mgl

ucos

e,gl

ycer

olor

�85

%in

osit

olF

reez

e-dr

ying

0%C

arpe

nter

etal

.,40

mgm

l−1

in1

mM

sodi

umbo

rate

(pH

7.8)

,P

FK

25m

MK

2SO

4,

2m

M(N

H4) 2

SO4

and

5m

M19

87di

thio

thre

itol

0%+

ZnS

O4

(up

to1

mM

)+

400

mM

treh

alos

eor

mal

tose

�80

%+

400

mM

sucr

ose

�60

%+

400

mM

gluc

ose

orga

lact

ose

0%+

0.9

mM

ZnS

O4

and

100

mM

treh

alos

eor

80%

mal

tose

+0.

9m

MZ

nSO

4an

d10

0m

Mgl

ucos

e90

%


Tab

le2

(Con

tinu

ed)

For

mul

atio

nco

mpo

siti

ons

Ref

eren

ces

Pro

tein

sSt

udy

cond

itio

nsR

emai

ning

Act

ivit

y(o

ras

stat

ed)

+0.

9m

MZ

nSO

4an

d10

0m

Mga

lact

ose

85%

50mg

ml−

1in

10m

Mpo

tass

ium

phos

phat

e0%

Car

pent

eran

dF

reez

e-dr

ying

PF

KC

row

e,19

89(p

H8.

0)an

d5

mM

dith

ioth

reit

ol�

45%

+15

0m

gm

l−1

treh

alos

e�

20%

+30

0m

gm

l−1

treh

alos

e15

0mg

ml−

1in

10m

Mpo

tass

ium

Car

pent

eran

dP

FK

Fre

eze-

dryi

ng0%

phos

phat

e(p

H8.

0)an

d5

mM

dith

ioth

reit

olC

row

e,19

89+

50m

gm

l−1

myo

-ino

sito

lF

reez

e-dr

ying

�20

%F

reez

e-dr

ying

+10

0m

gm

l−1

myo

-ino

sito

l�

5%


much more effective than sucrose in stabilizinglyophilized restriction enzyme PstI during storageat 37°C (Colaco et al., 1992). Nevertheless, therelative effect of the two disaccharides in stabiliz-ing solid proteins depends on many factors, suchas the relative concentration of the disaccharidesand the storage temperature at which stabilitystudies are conducted. Duddu and Dal Monte(1997) demonstrated that the aggregation rate fora lyophilized monoclonal antibody (2.5 mg ml−1)was about the same in the presence of either 31.3mg ml−1 sucrose or trehalose during storage at 5or 40°C for 2 months. However, at 60°C theaggregation percentage (6.0%) in sucrose formula-tion was significantly higher than that (1.1%) intrehalose formulation, apparently because thestorage temperature was higher than the glasstransition temperature of the sucrose formulation(59°C by DSC) but lower than that of the tre-halose formulation (80°C by DSC). Both sucroseand trehalose have been shown to stabilizelyophilized rFXIII and Humicola lanuginosa lipaseto a similar level during storage at 40°C for 3months, but sucrose was much less effective at60°C due to significant crystallization (Kreilgaardet al., 1998a, 1999). It has been shown that thetendency of crystallization of sucrose is higherthan that of trehalose during storage at tempera-tures above Tg (Hatley, 1997). Partly because ofthe difference in crystallization tendency, the glu-cose/trehalose (1:10, w/w) formulation preservedthe activity of G6PDH more effectively than theglucose/sucrose (1:10, w/w) formulation at storagetemperatures between 33 and 90.5°C (Sun andDavidson, 1998).

The level of protein stabilization afforded bysugars/polyols during long-term storage varies sig-nificantly. In certain cases, they destabilizeproteins, particularly at high concentrations. Afrequent cause of destabilization is crystallizationof these excipients during storage. Mannitol at100 mM does not improve the storage stability oflyophilized rFXIII at 40 or 60°C due to mannitolcrystallization (Kreilgaard et al., 1998a). A co-spray-dried formulation containing recombinanthumanized anti-IgE monoclonal antibody and10% (or 20%) mannitol was stable during storageat 5 or 30°C, but the antibody in 30% mannitol

formulation exhibited a drastic decrease in stabil-ity during storage due to mannitol crystallization(Costantino et al., 1998a). Costantino et al.(1998b) demonstrated that mannitol, sorbitol, ormethyl a-D-mannopyranoside inhibited aggrega-tion of lyophilized rhGH maximally at an excipi-ent:rhGH molar ratio of 131:1 during storage at50°C for 4 weeks. However, at higher ratios of300:1 and 1000:1, rhGH showed increased forma-tion of insoluble aggregates, apparently due tocrystallization of these excipients. In contrast, dis-accharides (lactose, trehalose, and cellobiose) weremore effective and stabilized the protein continu-ally at ratios of 131:1 and above. In a recentstudy, Tzannis and Prestrelski (1999a) demon-strated that the thermal stability (Tm) of trypsino-gen in a spray-dried sucrose formulation washighest at a mass ratio (sucrose:trypsinogen) of1:1 and varying the mass ratio in either directiondecreased the Tm.

Sugars/polyols may inhibit chemical degrada-tions in solid protein formulations. Sucrose at10% significantly decreased deamidation oflyophilized rhIL-1ra at 50°C (Chang et al.,1996a). Covalent dimerization of lyophilized in-sulin could be significantly suppressed in the pres-ence of trehalose (5 mg ml−1) at 35°C (Strickleyand Anderson, 1997).

5.3.3. PolymersMany polymers can increase the long-term sta-

bility of lyophilized proteins. Dextran, CMC,DEAE-dextran, and PEG have been shown toreduce aggregation of lyophilized BSA signifi-cantly during storage at 37°C (Liu et al., 1990).HP-b-CD was found to stabilize a mouse mono-clonal antibody during storage at 56°C (Ressinget al., 1992), to inhibit moisture-induced aggrega-tion of solid insulin (Katakam and Banga, 1995),to stabilize IL-2 against aggregation during stor-age at 5°C (Hora et al., 1992b), and to inhibit thedimerization of TNF during storage at 37°C(Hora et al., 1992a). Dextran 40 at 10% increasedthe activity of lyophilized elastase at 20 mg ml−1

(in 10 mM sodium acetate, pH 5.0) from 33 to82% during storage for 2 weeks at 40°C and 79%RH (Chang et al., 1993). Dextran (162 kD) at 3.5and 5% (w/v) improved the storage stability of


lyophilized rFXIII and Humicola lanuginosa li-pase, respectively, at 40 or 60°C (Kreilgaard et al.,1998a, 1999). Both PVPs and maltodextrin stabi-lized lyophilized invertase during incubation at90°C (Schebor et al., 1996; Cardona et al., 1997).Polyethyleneimine was shown to increase the stor-age stability of lyophilized LDH in a concentra-tion-dependent manner at 36°C (Table 2).

However, polymers may not always stabilizesolid proteins and, in certain cases, have adverseeffect, as discussed in Section 3.1. For example,neither insulin nor dextran could stabilizelyophilized restriction enzyme PstI during storageat 37°C (Colaco et al., 1992). Inclusion of dextran40 in a lyophilized IL-6 formulation containingsucrose significantly increased protein aggregationduring storage at 40°C for 9 months (Lueckel etal., 1998b). The destabilization can be attributedto a failure of inflexible dextran molecules tointeract with the protein effectively by hydrogenbonding. Apparently for the same reason, theactivity of lyophilized BO in a dextran formula-tion decreased faster than that in a PVA formula-tion during storage at 70°C (Nakai et al., 1998).

5.3.4. SaltsSalts have been shown to stabilize proteins in a

few cases. Liu et al. (1990) found that NaCl orsodium phosphate could significantly inhibit ag-gregation of lyophilized BSA (in water) on incu-bation at 37°C. rHA co-lyophilized with NaCl ata NaCl:protein weight ratio of 1:6 did not aggre-gate upon incubation at 37°C and 96% RH for 4days, while the protein without NaCl lost over80% solubility in just one day under the sameconditions (Costantino et al., 1995b). NaCl at anexcipient:protein weight ratio of 1:5 was also ableto reduce aggregation of lyophilized TT duringstorage at 37°C and 86% RH for 6 days (Schwen-deman et al., 1995).

5.3.5. SurfactantsAlthough surfactants may be effective protein

stabilizers during lyophilization, they seem to beincapable of stabilizing proteins effectively duringlong-term storage, based on a limited number ofstudies. Bush et al. (1998) demonstrated thatTween 80 could inhibit aggregation of FIX during

freezing and thawing, but did not provide enoughprotection for the lyophilized product during stor-age. Inclusion of 0.002% Tween 20 in alyophilized rFXIII formulation did not improveits storage stability at 40 and 60°C (Kreilgaard etal., 1998a). Although Tween 80 at 0.05 or 0.1%could inhibit spray-drying inactivation of LDH, itactually destabilized the dried protein during stor-age at 25, 40 and 60°C (Adler and Lee, 1999).

5.3.6. Miscellaneous compoundsA combination of 2.0% arginine and 2.3% car-

nitine significantly decreased aggregation oflyophilized IL-2 during storage at 37°C for 4weeks (Hora et al., 1992b). Combined use ofphenylalanine, arginine, and a mineral acid inhib-ited aggregation of vacuum-dried rhG–CSF orLDH during storage at 40°C (Mattern et al.,1999). Several excipients, such as D-glucaric acidand D-gluconic acid, have been shown to inhibitaggregation of lyophilized albumin during storage(Costantino et al., 1995b).

Calcium ions has been shown to protect solidrhDNase significantly against aggregation duringstorage at 40°C (Chen et al., 1999).

5.4. Formulation of solid protein pharmaceuticals

Although significant progress has been made inthe past decade in protein formulation, there isstill no single pathway to follow in formulating asolid protein product. In most cases, solid proteinproducts have been developed on a trial-and-errorbasis.

To achieve successful formulation of solidprotein products by lyophilization, one or more ofthe following formulation excipients may beneeded: a buffering agent(s), a bulking agent(s), aprotein stabilizer(s), and an antimicrobialagent(s), although stable proteins, such as recom-binant human tPA, may be lyophilized withoutany of the above agents (Overcashier et al., 1997).

5.4.1. Selection of a formulation pHLyophilization starts with preparation of a

protein solution. Many proteins in solution arestable only in a narrow pH range (see Section2.2). In addition, pH can strongly affect the solu-


bility of certain proteins, such as rIFN-b-1b (Be-taseron®), a hydrophobic protein (Lin et al.,1996). Therefore, an optimum pH is needed tokeep a protein stable and soluble in solution.

The solution pH has also been shown to affectstability of a protein during freezing and/orfreeze-drying. Freeze-thawing ovalbumin at neu-tral pH did not cause denaturation, but inducedsignificant structural changes at pH 1.9 (Koseki etal., 1990). Freezing human growth hormone(hGH) at pH 7.4 resulted in the formation ofmore insoluble aggregates than at pH 7.8 (Eck-hardt et al., 1991). Lyophilization of IL-2 at pH 7induced significant irreversible protein unfoldingand aggregation while at pH below 5, the proteinremained essentially native (Prestrelski et al.,1995). Thus, the solution pH must be optimal tominimize protein denaturation duringlyophilization.

On top of these effects, the formulation pHmay have significant impact on long-term stabilityof solid protein pharmaceuticals, as discussed inSection 5.2. Therefore, the formulation pH shouldbe optimal to allow maximum long-term stabilityfor lyophilized proteins.

To meet all these requirements in differentstages, the formulation pH needs to be carefullychosen. Proper selection of a solution pH is thefirst step toward stabilization of solid proteinpharmaceuticals. Very often, the most stable pHfor proteins in solution does not offer the beststability in solid state. This is because the inacti-vation mechanisms of proteins may well be differ-ent in the two different states, as reported forbovine pancreatic RNase A (Townsend andDeLuca, 1990). In such cases, a balanced pH mustbe used.

5.4.2. Selection of a buffering agent(s)Many buffering agents covering a wide pH

range are available for selection in formulatingsolid proteins. These agents include acetate, cit-rate, glycine, histidine, phosphate, Tris, etc. Abuffering agent, that also stabilizes a protein, ispreferable, such as histidine for freeze-dried FVIIISQ (O8 sterberg et al., 1997). For pH-sensitiveproteins, sodium phosphate should be avoidedbecause the selective crystallization of Na2HPO4

can cause a significant pH drop during freezing,denaturing proteins (see Section 2.2). Instead,potassium phosphate, citrate, histidine, and Triscan be used due to their minimal changes in pHduring freezing (Franks, 1990; Carpenter et al.,1997). In addition, decreasing the buffer concen-tration can also mitigate the pH shift (Pikal,1999).

Since the effect of different buffering agents onlong-term stability of lyophilized proteins is usu-ally unpredictable (see Section 5.2), selection of abuffering agent(s) can only rely on stability stud-ies. In addition, the selection of a proper bufferconcentration is also important, as the bufferconcentration not only affects the storage stabilityof lyophilized proteins (see Section 5.2) but alsoplays a critical role in stabilizing proteins duringlyophilization. For example, mannitol could notprotect b-galactosidase (2 mg ml−1) in water dur-ing lyophilization due to mannitol crystallization(Izutsu et al., 1993). In the presence of 10 mMsodium phosphate buffer (pH 7.4), crystallizationof mannitol at 50 mM was inhibited, and about95% of the enzyme activity was protected. In-creasing the concentration of sodium phosphatebuffer (pH 7.4) to 200 mM completely inhibitedthe crystallization of mannitol (500 mM) in b-galactosidase (2 mg ml−1) solution duringlyophilization (Izutsu et al., 1993).

5.4.3. Selection of a bulking agent(s)A crystallizing bulking agent(s) is usually

needed in a solid protein formulation to have oneor more of the following functions: to providemechanical support of the final cake, to improveproduct elegance, to improve formulation dissolu-tion, and to prevent product collapse and blow-out. A bulking agent(s) should have enoughsolubility, compatibility with the protein, no orminimal toxicity, and high eutectic temperature,allowing efficient freeze-drying.

Different bulking agents may affect stability ofsolid proteins to different degrees. Therefore,careful selection of a suitable bulking agent maybe necessary. For example, both the degradationand aggregation rates of lyophilized IL-1ra (pH6.5) during storage at 8, 30 or 50°C have beenshown to be different in three bulking agents:


mannitol, glycine or alanine (Chang et al., 1996c).Among these three agents, glycine was apparentlybest in protecting protein stability.

Two bulking agents are frequently used: glycineand mannitol. Glycine has several advantages,including non-toxicity, high solubility, and higheutectic temperature (Akers et al., 1995). Al-though both free or salt form of glycine can beused, it has been found that neutral glycine crys-tallizes rapidly, whereas glycine hydrochloridecrystallizes slowly, even at the same pH (Akers etal., 1995). As a bulking agent, mannitol maystabilize certain proteins. It has been used both asa bulking agent and as a stabilizer in a lyophilizedformulation of transforming growth factor-b1

(TGF-b1) (Gombotz et al., 1996).Most amino acids are potential bulking agents

as they easily crystallize out (Mattern et al., 1999).However, formation of acid salts reduces theirtendency to crystallize (Mattern et al., 1999). An-other crystalline agent, NaCl, is not preferable asa bulking agent due to its low eutectic and glasstransition temperature (Carpenter et al., 1997). Itwas, however, used as a bulking agent in alyophilized factor VIII SQ formulation because itsolubilized the protein (O8 sterberg et al., 1997).

An amorphous excipient(s) in a protein formu-lation may inhibit crystallization of the bulkingagent(s), thus affecting protein stability. Increas-ing the relative amount of sucrose in a mixture ofsucrose and glycine gradually inhibited crystalliza-tion of glycine, and at a weight ratio of 1.4:1(glycine:sucrose), crystallization of glycine wasnot detectable (Lueckel et al., 1998a). In a differ-ent study, crystallization of glycine was not ob-servable under microscope when its concentrationwas below 29% in the same mixture during freeze-drying; partial crystallization was observed at43%, but a good lyophilized cake was obtainedonly at a glycine concentration of about 50% orhigher (Kasraian et al., 1998). Similarly, mannitoldid not crystallize in the presence of a secondnon-crystallizing component, such as sucrose, lac-tose, maltose, or trehalose until the mannitol con-centration was over about 30% (w/w) (Kim et al.,1998). Because of this effect, lyophilized rFIXformulation containing 2% (0.26 M) glycine and1% sucrose (as a stabilizer) had both excellent

cake appearance and protein stability, but thatcontaining 1.7% (0.23 M) glycine and 2% sucroseonly offered reasonable protein stability withcrumbly cake appearance (Bush et al., 1998).Therefore, proper selection of a suitable bulkingagent(s) and its relative amount is critical.

5.4.4. Selection of a stabilizer(s)A solid protein product should be stable during

storage at least above 0°C, and preferably underroom conditions. To achieve this goal, a proteinstabilizer(s) is usually needed to protect a proteinduring lyophilization and/or long-term storage.Based on the stabilization mechanisms discussedin Section 5.3, a stabilizer(s) for a solid proteinproduct should be at least partially amorphousand able to replace water, forming intimate hy-drogen bonds with the protein. Formation of anamorphous glassy state is considered to be aprerequisite, not a guarantee, for protein stability(Pikal et al., 1991; Skrabanja et al., 1994). Since adisordered amorphous material has a lower en-ergy barrier for dissolution than a structured crys-talline solid, use of an amorphous stabilizer(s) canalso lead to faster dissolution of a solid proteinproduct (Miller et al., 1997).

The widely used stabilizers in solid proteinproducts are sugars. Most sugars do not crystal-lize under normal operating conditions (Changand Randall, 1992). Reducing sugars are generallynot preferable due to the possibility of reactingwith dried proteins via the Maillard reaction.However, reducing sugars can be used if they arebetter stabilizers. For example, dextrose, a reduc-ing sugar, has been shown to inhibit moisture-in-duced aggregation of bovine insulin significantlyat 1:1 protein:excipient ratio, while trehalose didnot stabilize the protein to a significant level(Katakam and Banga, 1995).

Among sugars, sucrose seems to be the mostcommonly selected. Recently, trehalose started togain attention due to several advantages, as dis-cussed in Section 3.1. Their relative effect in stabi-lizing solid proteins for long-term storage isprotein-dependent (also see Section 5.3). Here aresome more examples. Sucrose was shown to bemore effective than trehalose in inhibiting chemi-cal degradation of lyophilized IL-1ra during stor-


age at 8, 30 and 50°C (Chang et al., 1996c), whiletrehalose at 30 mg ml−1 in lyophilized IL-6 for-mulation inhibited IL-6 aggregation more effec-tively than sucrose during storage at both 25 and40°C for 9 months (Lueckel et al., 1998b). Twopossible reasons have been offered why trehalosewas more effective, (1) trehalose formulation hada higher glass transition temperature; and (2)sucrose could be hydrolyzed to form two reducingsugars for the Maillard reaction. However, theprotein molecular mobility in a trehalose formula-tion was shown to be higher than that in a sucroseformulation at certain temperatures, which mightmake trehalose a less effective stabilizer than su-crose depending on the protein, degradationmechanisms, and storage conditions (Miller et al.,1997).

Certain salts can be used to stabilize solidproteins (also see Section 3.1). However, the pres-ence of uncrystallized salts in a freeze-concentrateusually depresses Tg% , so the salt content in proteinformulations should be kept to a minimum(Franks, 1990). Several factors may affect theextent of salt crystallization, including the natureof salt, salt concentration, and cooling rate.Chang and Randall (1992) have classified saltsinto three types based on their glass-forming ten-dency at a given cooling rate and subsequentthermal history, (1) crystallizing salts such asmaleic acid, Na2HPO4, Na2SO4, Na2CO3, KCl,and (NH4)2SO4; (2) partially crystallizing (doublyunstable glass) salts such as NaCl, NaHCO3,K2HPO4, KH2PO4, CaCl2, MgCl2, glycine, andb-alanine; and (3) glass-forming salts such asNaH2PO4, sodium/potassium citrate, citric acid,histidine, and sodium/potassium acetate. Sinceglass-forming excipients can inhibit salt crystal-lization, salts can be potential protein stabilizersin the presence of other amorphous excipients(Hatley and Franks, 1991).

Some polymers can be chosen as protein stabi-lizers in solid state, as they can increase Tg ofprotein formulations. On the other hand, theymay not be as effective as sugars due to inefficienthydrogen bonding with proteins (Schebor et al.,1996; Cardona et al., 1997; Kreilgaard et al.,1998a). An alternative is to use both polymersand sugars together to achieve enhanced protein

stabilization. While polymers increase the Tg of aprotein formulation and inhibit crystallization ofother excipients, sugars can form intimate hydro-gen bonds with proteins. This strategy, however,has not been very effective at least for someproteins. Inclusion of dextran 40 in a lyophilizedIL-6 formulation containing sucrose significantlyincreased protein aggregation during storage at40°C for 9 months (Lueckel et al., 1998b).

Proteins, like other polymers, have high Tgs.Therefore, increasing protein:excipient ratio in-creases Tg of protein formulations. Increasing thecontent of a lyophilized monoclonal antibodyfrom 2.5 to 25 mg ml−1 in sucrose (or trehalose at31.3 mg ml−1) formulation increased the Tg of theformulation from 59 (or 80) to 89 (or 100)°C byDSC (Duddu and Dal Monte, 1997). Increasingthe relative content of rbSt to 50% in a sucrose/rbSt mixture gradually increased the Tg from74°C for pure sucrose to 96°C for the mixture(Sarciaux and Hageman, 1997). Therefore, therelative protein concentration in a formulationshould be kept relatively high to prevent collapse.More importantly, a high protein:excipient ratiostrongly inhibits excipient crystallization (also seeSection 5.3). For example, the lyophilized rhD-Nase–mannitol formulations were partially crys-talline when the relative protein quantity was 17%(w/w) or less, but became amorphous at 80% ormore (Chan et al., 1999). At a higherprotein:sucrose ratio, lyophilized Humicola lanug-inosa lipase was shown to be more stable duringstorage at 60°C, due to effective inhibition ofsucrose crystallization by the protein (Kreilgaardet al., 1999). In a more detailed study, the Tcry ofsucrose at 4.4% water content was �70°C andincreased to 120°C when 20% rbSt was included(Sarciaux and Hageman, 1997). On the otherhand, increasing the protein concentration toohigh may eventually destabilize a protein due toinsufficient quantity of a stabilizer, as has beenreported for rhIL-1ra (Chang et al., 1996c).

The presence of a bulking agent may affectproperties of an amorphous excipient. The Tg% of amixture of sucrose and glycine (or lysine–HCl)decreased with increasing amino acid:sucrose ra-tio (Lueckel et al., 1998a). Addition of glycine (upto 71%, w/w) in a sucrose solution decreased


linearly the Tg% of the mixture from −32 to−52°C (Kasraian et al., 1998). When glycineconcentration was over 86%, no Tg% was observed.Similarly, increasing mannitol concentrations upto 30% decreased Tg% of sugars, such as sucrose,lactose, maltose, and trehalose (Kim et al., 1998).To keep Tg% at a level high enough for efficientlyophilization, the relative quantity of a bulkingagent should be properly chosen and optimized.

5.4.5. Selection of other excipientsOther formulation excipients, such as antimi-

crobial agents or solubilizers may be used depend-ing on the protein. Both Tween 80 and SDS(more effective) facilitate solubilization of rIFN-b-1b (Betaseron®) (Lin et al., 1996). SDS has alsobeen used to solubilize IL-2 in Proleukin® (Physi-cians’ Desk Reference, 1999).

5.4.6. O6erall consideration of formulationexcipients

Generally, the total quantity of solid in proteinformulations is between 2 and 10%. While a solidcontent lower than 2% may not form a strongcake, higher amount of solid (\10%) may bedifficult to process (Hatley et al., 1996; Carpenteret al., 1997; Jennings, 1999; Willemer, 1999). Inaddition, higher solid content may affect reconsti-tution of lyophilized formulations. Breen et al.(1998) demonstrated that lyophilized cakes pre-pared from 110 mg ml−1 bulk took approxi-mately 60 min to reconstitute as compared withless than 5 min for those from dilute bulks. SEManalysis showed that cakes lyophilized fromhighly concentrated bulks had smaller pores withthicker walls than those from more dilute bulks.

The relative quantity of different excipients isalso important because their physical propertiesare mutually affected. Excipients that providemore than one function in protein formulationsshould be selected first, such as sugars, which maybe used as both cryoprotectants and lyoprotec-tants. High levels of buffers or salts should beavoided because this may lead to potential pHchanges during freezing and likely depression inTg% and Tg of dried formulations (Pikal, 1990b).

To expedite selection of formulation excipients,two screening methods have been used: measure-

ment of Tg of a lyophilized formulation in thepresence of different excipients or determinationof IR spectrum of a formulation in comparison tothat of a reference (see Section 2.3). Both methodsshould be used with caution, as these parametersmay not always reflect changes in protein activityor stability.

Recent investigations suggest that combined useof sucrose as a protein stabilizer and glycine as abulking agent may be a good starting point whenformulating a solid protein product. The combi-nation of these two excipients has been usedsuccessfully in formulating several proteins. Therecently developed albumin-free formulation forrecombinant factor VIII (rFVIII) contains su-crose, glycine, histidine, CaCl2, and NaCl (Nayar,1998). Lyophilized rFIX formulation contains su-crose, glycine, histidine, and polysorbate 80 (Bushet al., 1998). Chang et al. (1996c) developed astable lyophilized IL-1ra formulation, which con-tains sucrose, glycine, and sodium citrate. Thisstable IL-1ra formulation was also able to stand asingle-step freeze-drying cycle (Chang and Fis-cher, 1995).

5.4.7. Selection of containers and stoppersIn close relation to selection of formulation

excipients, a compatible container should be used.Type I borosilicate glass (treated or untreated) isusually the material of choice for containers dueto its strong chemical resistance and low level ofleachables. Nevertheless, proteins can be absorbedto glass surfaces to different extents (Gombotz etal., 1996). Since the loss of protein activity fromglass surface absorption and surface-induced de-naturation is protein-dependent, containers needto be evaluated on a case-by-case basis (Burke etal., 1992).

The volume of a solution for bolus intravenousinjection generally does not exceed 10 ml. There-fore, the size of containers for these applicationsis usually smaller than 20 ml, assuming the vol-ume of the solution to be lyophilized is ideallybetween about 20 and 50% of the container vol-ume. A relatively small filling volume is preferablefor an efficient freeze-drying cycle and a higherprotein concentration often results in a lower lossof activity due to surface absorption and stability


during freeze-drying. The lyophilized product canbe reconstituted with different volumes of wateror solution based on the protein solubility andstability.

Container stoppers may significantly affectlong-term stability of solid protein pharmaceuti-cals. This effect is often due to a stopper-inducedincrease in moisture content of lyophilized formu-lations during storage. The percent change inmoisture content depends on both hygroscopicityand quantity of the protein formulation in thecontainer. Also, oxygen permeation through stop-pers or stopper leakage may increase the partialpressure of oxygen in a product vial and causepotential oxidation of proteins during storage(Wang et al., 1997).

Three stopper-related processes can lead to anincrease in residual moisture of protein formula-tions, (1) transfer of moisture from stoppers to theformulation; (2) diffusion or transmission ofmoisture through the stopper; and (3) microleaksin the stopper-vial seal (House and Mariner,1996). The first process is usually dominant (Fordand Dawson, 1994). Stoppers absorb moistureduring routine steam sterilization and sterilizedstoppers can transfer some of its moisture back tothe protein product (DeGrazio and Flynn, 1992;Corveleyn et al., 1997). In fact, the steam auto-claving process is responsible for the majority ofwater driven into elastomeric stoppers. Chang etal. (1996c) demonstrated that two out of three lotsof siliconized stoppers showed transfer of mois-ture to the lyophilized rhIL-1ra formulation, re-sulting in a significant increase in moisturecontent during stability studies. Storage of alyophilized IL-6 formulation at 25 or 40°C for 9months led to an increase in moisture contentfrom 0.5 to 2% (Lueckel et al., 1998b). The in-crease in moisture content can cause a significantdrop in Tg. It has been shown that moisturetransfer from the stopper dropped the Tg of asucrose-formulated rFXIII formulation by 11°Cafter storage at 40°C for 3 months (Kreilgaard etal., 1998a) and that of a vacuum-dried LDHformulation by 35°C after storage at 60°C for 26weeks (Mattern et al., 1999).

There are many types of stopper materialsavailable. Butyl or halobutyl rubber is often the

choice for solid protein formulations because theyhave relatively low moisture absorption and vaportransmission rates (DeGrazio and Flynn, 1992).Among halobutyl stoppers, bromobutyl stoppershave been shown to be more resistant thanchlorobutyl stoppers to moisture absorption dur-ing storage and steam sterilization (Corveleyn etal., 1997). To minimize the moisturizing effect,stoppers can be autoclaved at 121°C for 30 minand immediately dried in an oven at 135°C for 5h before use. On the other hand, overdried stop-pers may absorb moisture from the lyophilizedformulation and dropped the moisture content ofthe formulation during storage (Corveleyn et al.,1997). The integrity of stoppers can be testedunder high humidity by monitoring the change inmoisture content of the protein formulation.Therefore, stoppers should be carefully chosenbased on their compatibility with the protein for-mulation, resistance to formulation pH, excipientsand sterilization, moisture/vapor transfer prop-erty, and resealability.

5.4.8. Stability testing of final formulationsSelection of the final protein formulation can

only be based on stability studies. A variety offormulation parameters in stability studies can beused to compare different formulations, includingprotein activity, cake physical attributes (shape,color and texture), particulate formation, mois-ture content, and reconstitution (or rehydratabil-ity). Another parameter is the volume of proteinformulation preparation, which should notchange much after lyophilization and during stor-age. Although minor shrinkage in volume may beseen, especially at low solute concentrations(Daukas and Trappler, 1998), a significant drop involume may be an indication of formulation melt-back or collapse.

To expedite selection of the final protein formu-lation, accelerated stability studies are frequentlyconducted. These stressed stability conditions in-clude high temperature, high humidity, or inten-sive lighting. One key issue in conductingaccelerated stability studies is whether and howwell the data obtained at high temperatures canbe extrapolated to those under real-time condi-tions. Often, protein stability results obtained at


high temperatures do not reflect or predict whathappens under real-time conditions, due to thecomplexity of multiple protein degradation path-ways at different temperatures. For example, Sunet al. (1998) found that the temperature depen-dence of G6PDH inactivation in carbohydrate(glucose or sucrose) glassy matrix deviated fromArrhenius relationship between 8 and 93°C. Onthe other hand, if the multiple degradation pro-cesses in proteins can be described separately, orthe rate-limiting degradation step does not changewithin a certain temperature range, prediction ofprotein stability based on accelerated stabilitystudies is very optimistic. The loss of lactase activ-ity in a lyophilized PVP formulation followedArrhenius behavior well in a temperature rangebetween 37 and 70°C (Mazzobre et al., 1997).Yoshioka et al. (1994) examined the stability ofsix protein preparations, including a-chy-motrypsin troche, a-chymotrypsin tablet, brome-lain tablet A and B, kallikrein capsule, andb-galactosidase powder at elevated temperaturesbetween 40 and 70°C (50 or 75% RH). Theyfound that all of them had complex kinetics, butthe inactivation rate exhibited approximately lin-ear Arrhenius relationship. Nevertheless, forproteins with unknown degradation pathways,real-time stability testing has to be conducted forselection of the final protein formulation.

6. Summary

Lyophilization (freeze-drying) is the most com-mon process for making solid protein pharmaceu-ticals. However, this process can generate avariety of stresses, and denature proteins to vari-ous degrees. These stresses include low tempera-ture, formation of dendritic ice crystals, increasein ionic strength, pH changes, phase separation,and removal of the protein hydration shell. Struc-tural changes in proteins during lyophilization canbe either reversible or irreversible depending onthe protein, and can be conveniently monitoredby IR. Proteins sensitive to freezing and/or dryingstresses can be stabilized by selection of propercryo- and/or lyoprotectants. The commonly usedcryoprotectants include sugars/polyols, non-

aqueous solvents, polymers, protein itself, surfac-tants, and amino acids. These cryoprotectants canalso be used as potential lyoprotectants exceptnon-aqueous solvents. These stabilizers may pro-tect proteins by one or more of the followingmechanisms: preferential interaction, replacementof water, formation of a glass, hydrogen bonding,and steric hindrance.

The design of a lyophilization cycle may havesignificant impact on protein stability during andafter lyophilization. Lyphilization cycle-relatedstresses include freezing rate and temperature,thermal treatment condition, drying rate and tem-perature, and the final moisture content. There-fore, efforts should be made to design a cycle,which is robust and efficient, and has minimaladverse effects on protein stability. To design sucha cycle, protein formulations need thorough char-acterization and all critical temperatures need tobe determined as discussed in Section 4.1.

Solid protein pharmaceuticals may experience avariety of instabilities. Major instabilities includeaggregation, deamidation, oxidation, the Maillardreaction, hydrolysis, and disulfide bond forma-tion/exchange. Many factors affect these instabili-ties, including storage temperature, glasstransition temperature of the formulation, resid-ual moisture content, formulation ‘pH’, crystal-lization of amorphous excipients, and presence ofdestabilizing excipients or contaminants. Theseinstabilities may be minimized by proper selectionof formulation pH, residual moisture content, andmore importantly, formulation stabilizers, such assugars/polyols, polymers, salts, and surfactants.

Formulation of solid protein pharmaceuticalsmay require not only suitable stabilizers but alsoother excipients such as bulking, buffering, andantimicrobial agents. Since the physical propertiesof these agents are mutually affected, the relativequantity of these agents in a protein formulationis critical and should be determined based onsound experimentation and appropriate stabilitystudies. A proposed protein formulation for initialdevelopment trial is 2% glycine, 1% sucrose, and20 mM buffering agent controlling the formula-tion pH. To expedite selection of formulationexcipients, two screening methods may be usedwith caution: comparison of Tgs or IR spectra of


formulations in the presence of different excipients.However, selection of the final protein formulationrequires real-time stability studies.

In summary, development of a lyophilizedprotein product usually takes an enormous amountof time, labor, and effort, simply because there isno single, short, and mature pathway to follow informulating such a product, and many experimentsare done on a trial-and-error basis. This trend willcontinue until a breakthrough is achieved in under-standing the basic behavior of proteins and theirstabilization.

7. Abbreviations

aFGF acidic fibroblast growth factorbasic fibroblast growth factorbFGFbovine g-globulinBGGbilirubin oxidaseBObovine pancreatic trypsin inhibitorBPTI

BSA bovine serum albumin3-[(3-cholamidepropyl)CHAPS-dimethylammonio]-1-propanesulfatecarboxymethyl celluloseCMCcytochrome cCyt c

DMF dimethylformamidedimethylsulphoxideDMSOdifferential scanning calorimetryDSCgranulocyte colony-stimulatingG-CSFfactor

GDH glutamate dehydrogenaseG6PDH glucose-6-phosphate

dehydrogenasegranulocyte macrophage colonyGM-CSFstimulating factor

HEC hydroxyethyl cellulosehIGF-I human insulin-like growth factor I

Humicola lanuginosa lipaseHLLhydroxypropyl-b-cyclodextrinHP-b-CDhydroxypropyl methylcelluloseHPMChuman serum albuminHSAinterferon-bIFN-binterferon-gIFN-g

IL-1a interleukin-1ainterleukin-1bIL-1b

IL-2 interleukin-2

infrared spectroscopyIRlactate dehydrogenaseLDHlow molecular weight urokinaseLMW-UKmyoglobinMbmaleate dehydrogenaseMDHmethionyl human growth hormoneMet-hGHmass spectroscopyMS

NMR nuclear magnetic resonancespectroscopy

PEG polyethylene glycolphosphofructokinasePFKpolyvinylalcoholPVApolyvinylpyrrolidonePVPrecombinant bovine somatotropinrbStrecombinant consensus a-rconIFNinterferon

rFIX recombinant factor IXrFXIII recombinant factor XIII

recombinant human albuminrHArhCNTF recombinant human ciliary neuro-

trophic factorrecombinant human interleukin-1rhIL-1rareceptor antagonist

rhKGF recombinant human keratinocytegrowth factorrecombinant human macrophagerhMCSFcolony-stimulating factor

RNase A ribonuclease ASDS sodium dodecyl sulfate

transforming growth factorTGFtumor necrosis factor bindingTNFbpprotein

tPA tissue plasminogen activatorTT tetanus toxoid

Acknowledgements

This review would not be possible without thesupport and encouragement of Drs Rajiv Nayar,Sheryl Martin-Moe, and Bob Kuhn. I am alsoindebted to Drs Rajiv Nayar, Jeff Hey, XingHangMa, and especially, the two referees for their criticalreview of the manuscript and their valuable com-ments. The careful editing of the manuscript byBlanca Martinez is also acknowledged


References

Adams, G.D.J., Ramsay, J.R., 1996. Optimizing the lyophiliza-tion cycle and the consequences of collapse on the pharma-ceutical acceptability of Erwinia L-asparaginase. J. Pharm.Sci. 85, 1301–1305.

Adler, M., Lee, G., 1999. Stability and surface activity of lactatedehydrogenase in spray-dried trehalose. J. Pharm. Sci. 88,199–208.

Akers, M.J., Milton, N., Byrn, S.R., Nail, S.L., 1995. Glycinecrystallization during freezing: the effect of salt form, pH,and ionic strength. Pharm. Res. 12, 1457–1461.

Alden, M., Magnusson, A., 1997. Effect of temperature historyon the freeze-thawing process and activity of LDH formu-lations. Pharm. Res. 14, 426–430.

Aldous, B.J., Auffret, A.D., Franks, F., 1995. The crystalliza-tion of hydrate from amorphous carbohydrates. Cryo-Let-ters 16, 181–186.

Allison, S.D., Dong, A., Carpenter, J.F., 1996. Counteractingeffects of thiocyanate and sucrose on chymotrypsinogensecondary structure and aggregation during freezing, dryingand rehydration. Biophys. J. 71, 2022–2032.

Allison, S.D., Randolph, T.W., Manning, M.C., Middleton, K.,Davis, A., Carpenter, J.F., 1998. Effects of drying methodsand additives on structure and function of actin: mechanismsof dehydration-induced damage and its inhibition. Arch.Biochem. Biophy. 358, 171–181.

Allison, S.D., Chang, B., Randolph, T.W., Carpenter, J.F.,1999. Hydrogen bonding between sugar and protein isresponsible for inhibition of dehydration-induced proteinunfolding. Arch. Biochem. Biophys. 365, 289–298.

Anchordoquy, T.J., Carpenter, J.F., 1996. Polymers protectlactate dehydrogenase during freeze-drying by inhibitingdissociation in the frozen state. Arch. Biochem. Biophys.332, 231–238.

Andersson, M.M., Hatti-Kaul, R., 1999. Protein stabilizingeffect of polyethyleneimine. J. Biotechnol. 72, 21–31.

Andya, J.D., Maa, Y.-F., Costantino, H.R., Nguyen, P.-A.,Dasovich, N., Sweeney, T.D., Hsu, C.C., Shire, S.J., 1999.The effect of formulation excipients on protein stability andaerosol performance of spray-dried powders of a recombi-nant humanized anti-IgE monoclonal antibody. Pharm. Res.16, 350–358.

Angell, C.A., 1995. Formation of glasses from liquid andbiopolymers. Science 267, 1924–1935.

Arakawa, T., Yoshiko, K., Carpenter, J.F., 1991. Protein-sol-vent interactions in pharmaceutical formulations. Pharm.Res. 8, 285–291.

Arakawa, T., Prestrelski, S.J., Kenney, W.C., Carpenter, J.F.,1993. Factors affecting short-term and long-term stabilitiesof proteins. Adv. Drug Deliv. Rev. 10, 1–28.

Bam, N.B., Randolph, T.W., Cleland, J.L., 1995. Stability ofprotein formulations: investigation of surfactant effects bya novel EPR spectroscopic technique. Pharm. Res. 12, 2–11.

Bell, L.N., Hageman, M.J., Muraoka, L.M., 1995. Thermallyinduced denaturation of lyophilized bovine somatotropinand lysozyme as impacted by moisture and excipients. J.Pharm. Sci. 84, 707–712.

Bell, L.N., Hageman, M.J., 1996. Glass transition explanationfor the effect of polyhydroxy compounds on protein denat-uration in dehydrated solids. J. Food Sci. 61, 372–378.

Belton, P.S., Gil, A.M., 1994. IR and Raman spectroscopicstudies of the interaction of trehalose with hen egg whitelysozyme. Biopolymers 34, 957–961.

Bindschaedler, C., 1999. Lyophilization process validation. In:Rey, L., May, J.C. (Eds.), Freeze-Drying/Lyophilization ofPharmaceutical and Biological Products, vol. 96. MarcelDekker, New York, pp. 373–408.

Brange, J., Havelund, S., Hougaard, P., 1992a. Chemicalstability of insulin. 2. Formation of higher molecular weighttransformation products during storage of pharmaceuticalpreparations. Pharm. Res. 9, 727–734.

Brange, J., Langkjaer, L., Havelund, S., Vølund, A., 1992b.Chemical stability of insulin. 1. Hydrolytic degradationduring storage of pharmaceutical preparations. Pharm. Res.9, 715–726.

Breen, E.D., Costantino, H.R., Hsu, C.C., Shire, S.J., 1998.Affect of bulk concentration on reconstitution of lyophilizedprotein formulations. Pharm. Sci. 1 (Suppl.), S540.

Buera, M.D.P., Chirife, J., Resnik, S.L., Wetzler, G., 1987.Nonenzymic browning in liquid model systems of high wateractivity: kinetics of color changes due to Maillard’s reactionbetween different single sugars and glycine and comparisonwith caramelization browning. J. Food Sci. 52, 1063–1067.

Buera, M.P., Levi, G., Karel, M., 1992. Glass transition inpolyvinylpyrrolidone: effect of molecular weight and dilu-ents. Biotechnol. Prog. 8, 144–148.

Bunk, D.M., 1997. Characterization of the glycation of albuminin freeze-dried and frozen human serum. Anal. Chem. 69,2457–2463.

Burke, C.J., Steadman, B.L., Volkin, D.B., Tsai, P.-K., Bruner,M.W., Middaugh, C.R., 1992. The adsorption of proteinsto pharmaceutical container surfaces. Int. J. Pharm. 86,89–93.

Bush, L., Webb, C., Bartlett, L., Burnett, B., 1998. Theformulation of recombinant factor IX: stability, robustness,and convenience. Semin. Hematol. 35, 18–21.

Cardona, S., Schebor, C., Buera, M.P., Karel, M., Chirife, J.,1997. Thermal stability of invertase in reduced-moistureamorphous matrices in relation to glassy state and trehalosecrystallization. J. Food Sci. 62, 105–112.

Careri, G., Giansanti, A., Gratton, E., 1979. Lysozyme filmhydration events: an IR and gravimetric study. Biopolymers18, 1187–1203.

Carpenter, J.F., Crowe, J.H., 1989. An infrared spectroscopicstudy of the interactions of carbohydrates with driedproteins. Biochemistry 28, 3916–3922.

Carpenter, J.F., Hand, S.C., Crowe, L.M., Crowe, J.H., 1986.Cryoprotection of phosphofructokinase with organic so-lutes: characterization of enhanced protection in the pres-ence of divalent cations. Arch. Biochem. Biophys. 250,505–512.


Carpenter, J.F., Crowe, L.M., Crowe, J.H., 1987. Stabilizationof phosphofructokinase with sugars during freeze-drying:characterization of enhanced protection in the presence ofdivalent cations. Biochim. Biophys. Acta 923, 109–115.

Carpenter, J.F., Crowe, J.H., Arakawa, T., 1990. Comparisonof solute-induced protein stabilization in aqueous solutionand in frozen and dried state. J. Dairy Sci. 73, 3627–3636.

Carpenter, J.F., Arakawa, T., Crowe, J.H., 1991. Interactionsof stabilizing additives with proteins during freeze-thawingand freeze-drying. Dev. Biol. Stand. 74, 225–239.

Carpenter, J.F., Prestrelski, S.J., Arakawa, T., 1993. Separa-tion of freezing and drying-induced denaturation oflyophilized proteins using stress-specific stabilization. I.Enzymatic activity and calorimetric studies. Arch.Biochem. Biophys. 303, 456–464.

Carpenter, J.F., Pikal, M.J., Chang, B.S., Randolph, T.W.,1997. Rational design of stable lyophilized protein formu-lations: some practical advice. Pharm. Res. 14, 969–975.

Carpenter, J.F., Prestrelski, S.J., Dong, A., 1998. Applicationof infrared spectroscopy to development of stablelyophilized protein formulations. Eur. J. Pharm. Bio-pharm. 45, 231–238.

Carpenter, J.F., Izutsu, K., Randolph, T.W., 1999. Freezing-and drying-induced perturbations of protein structure andmechanisms of protein protection by stabilizing additives.In: Rey, L., May, J.C. (Eds.), Freeze-Drying/Lyophiliza-tion of Pharmaceutical and Biological Products, vol. 96.Marcel Dekker, New York, pp. 123–160.

Chan, H.K., Gonda, I., 1998. Solid state characterization ofspray-dried powders of recombinant human deoxyribonu-clease (rhDNase). J. Pharm. Sci. 87, 647–654.

Chan, H.K., Au-Yeung, K.L., Gonda, I., 1999. Developmentof a mathematical model for the water distribution infreeze-dried solids. Pharm. Res. 16, 660–665.

Chang, B.S., Randall, C.S., 1992. Use of subambient thermalanalysis to optimize protein lyophilization. Cryobiology29, 632–656.

Chang, B.S., Randall, C.S., Lee, Y.S., 1993. Stabilization oflyophilized porcine pancreatic elastase. Pharm. Res. 10,1478–1483.

Chang, B.S., Fischer, N.L., 1995. Development of an efficientsingle step freeze drying cycle for protein formulations.Pharm. Res. 12, 831–837.

Chang, B.S., Beauvais, R.M., Dong, A., Carpenter, J.F.,1996a. Physical factors affecting the storage stability offreeze-dried interleukin-1 receptor antagonist: glass transi-tion and protein conformation. Arch. Biochem. Biophys.331, 249–258.

Chang, B.S., Kendrick, B.S., Carpenter, J.F., 1996b. Surface-induced denaturation of proteins during freezing and itsinhibition by surfactants. J. Pharm. Sci. 85, 1325–1330.

Chang, B.S., Reeder, G., Carpenter, J.F., 1996c. Developmentof a stable freeze-dried formulation of recombinant humaninterleukin-1 receptor antagonist. Pharm. Res. 13, 243–249.

Chen, B., Costantino, H.R., Liu, J., Hsu, C.C., Shire, S.J.,1999. Influence of calcium ions on the structure and stabil-

ity of recombinant human deoxyribonuclease I in theaqueous and lyophilized states. J. Pharm. Sci. 88, 477–482.

Chuyen, N.V., 1998. Maillard reaction and food processing.Application aspects. Adv. Exp. Med. Biol. 434, 213–235.

Cleland, J.L., Powell, M.F., Shire, S.J., 1993. The developmentof stable protein formulations: a close look at proteinaggregation, deamidation, and oxidation. Crit. Rev. Ther.Drug Carrier Syst. 10, 307–377.

Colaco, C., Sen, S., Thangavelu, M., Pinder, S., Roser, B.,1992. Extraordinary stability of enzymes dried in trehalose:simplified molecular biology. Bio. Technol. 10, 1007–1011.

Cooper, E.A., Knutson, K., 1995. Fourier transform infraredspectroscopy investigations of protein structures. In: Her-ron, J.N., Jiskoot, W., Crommelin, D.J.A. (Eds.), PhysicalMethods to Characterize Pharmaceutical Proteins. PlenumPress, New York, pp. 101–143.

Corveleyn, S., De Smedt, S., Remon, J.P., 1997. Moistureabsorption and desorption of different rubber lyophiliza-tion closures. Int. J. Pharm. 159, 57–65.

Costantino, H.R., Langer, R., Klibanov, A.M., 1994a. Solid-phase aggregation of proteins under pharmaceutically rele-vant conditions. J. Pharm. Sci. 83, 1662–1669.

Costantino, H.R., Langer, R., Klibanov, A.M., 1994b. Mois-ture-induced aggregation of lyophilized insulin. Pharm.Res. 11, 21–29.

Costantino, H.R., Griebenow, K., Mishra, P., Langer, R.,Klibanov, A.M., 1995a. Fourier-transform infrared spec-troscopic investigation of protein stability in thelyophilized form. Biochim. Biophy. Acta 1253, 69–74.

Costantino, H.R., Langer, R., Klibanov, A.M., 1995b. Aggre-gation of a lyophilized pharmaceutical protein, recombi-nant human albumin: effect of moisture and stabilizationby excipients. Biotechnology 13, 493–496.

Costantino, H.R., Schwendeman, S.P., Griebenow, K.,Klibanov, A.M., Langer, R., 1996. The secondary struc-ture and aggregation of lyophilized tetanus toxoid. J.Pharm. Sci. 85, 1290–1293.

Costantino, H.R., Andya, J.D., Nguyen, P.-A., Dasovich, N.,Sweeney, T.D., Shire, S.J., Hsu, C.C., Maa, Y.-F., 1998a.Effect of mannitol crystallization on the stability and aero-sol performance of a spray-dried pharmaceutical protein,recombinant humanized anti-IgE monoclonal antibody. J.Pharm. Sci. 87, 1406–1411.

Costantino, H.R., Carrasquillo, K.G., Cordero, R.A., Mu-menthaler, M., Hsu, C.C., Griebenow, K., 1998b. Effect ofexcipients on the stability and structure of lyophilizedrecombinant human growth hormone. J. Pharm. Sci. 87,1412–1420.

Costantino, H.R., Curley, J.G., Wu, S., Hsu, C.C., 1998c.Water sorption behavior of lyophilized protein–sugar sys-tems and implications for solid-state interactions. Int. J.Pharm. 166, 211–221.

Costantino, H.R., Schwendeman, S.P., Langer, R., Klibanov,A.M., 1998d. Deterioration of lyophilized pharmaceuticalproteins. Biochemistry (Moscow) 63, 357–363.

Crowe, J.H., Hoekstra, F.A., Crowe, L.M., 1992. Anhydro-biosis. Annu. Rev. Physiol. 54, 579–599.


Crowe, J.H., Crowe, L.M., Carpenter, J.F., 1993a. Preservingdry biomaterials: the water replacement hypothesis. Bio-pharm. 6 (1), 28–37.

Crowe, J.H., Crowe, L.M., Carpenter, J.F., 1993b. Preservingdry biomaterials: the water replacement hypothesis. Bio-pharm. 6 (2), 40–43.

Crowe, L.M., Reid, D.S., Crowe, J.H., 1996. Is trehalosespecial for preserving dry biomaterials? Biophy. J. 71,2087–2093.

Crowe, J.H., Carpenter, J.F., Crowe, L.M., 1998. The role ofvitrification in anhydrobiosis. Annu. Rev. Physiol. 60, 73–103.

Darcy, P., Buckton, G., 1997. The influence of heating/dryingon the crystallization of amorphous lactose after structuralcollapse. Int. J. Pharm. 158, 157–164.

Daukas, L.A., Trappler, E.H., 1998. Assessing the quality oflyophilized parenterals. Pharm. Cosmetic Quality 2, 21–25.

Dawson, P.J., 1992. Effect of formulation and freeze-drying onthe long-term stability of rDNA-derived cytokines. Dev.Biol. Stand. 74, 273–282.

DeGrazio, F., Flynn, K., 1992. Lyophilization closures forprotein based drugs. J. Parenter. Sci. Technol. 46, 54–61.

De Young, L.R., Dill, K.A., Fink, A.L., 1993. Aggregationand denaturation of apomyoglobin in aqueous urea solu-tions. Biochemistry 32, 3877–3886.

Dill, K.A., Alonso, D.O.V., Hutchinson, K., 1989. Thermalstabilities of globular proteins. Biochemistry 28, 5439–5449.

Dill, K.A., 1990. Dominant forces in protein folding. Bio-chemistry 29, 7133–7155.

Dong, A., Prestrelski, S.J., Allison, S.D., Carpenter, J.F.,1995. Infrared spectroscopic studies of lyophilization- andtemperature-induced protein aggregation. J. Pharm. Sci.84, 415–423.

Duddu, S.P., Dal Monte, P.R., 1997. Effect of glass transitiontemperature on the stability of lyophilized formulationscontaining a chimeric therapeutic monoclonal antibody.Pharm. Res. 14, 591–595.

Duddu, S.P., Zhang, G., Dal Monte, P.R., 1997. The relation-ship between protein aggregation and molecular mobilitybelow the glass transition temperature of lyophilized for-mulations containing a monoclonal antibody. Pharm. Res.14, 596–600.

Eckhardt, B.M., Oeswein, J.Q., Bewley, T.A., 1991. Effect ofFreezing on aggregation of human growth hormone.Pharm. Res. 8, 1360–1364.

Fan, J., Cooper, E.I., Angell, C.A., 1994. Glasses with strongcalorimetric b-glass transitions and the relation to theprotein glass transition problem. J. Phys. Chem. 98, 9345–9349.

Forbes, R.T., Davis, K.G., Hindle, M., Clarke, J.G., Maas, J.,1998. Water vapor sorption studies on the physical stabil-ity of a series of spray-dried protein/sugar powders forinhalation. J. Pharm. Sci. 87, 1316–1321.

Ford, A.W., Dawson, P.J., 1994. Effect of type of container;storage temperature and humidity on the biological activityof freeze-dried alkaline phosphatase. Biologicals 22, 191–197.

Fox, K.C., 1995. Putting proteins under glass. Science 267,1922–1923.

Franks, F., 1990. Freeze-drying: from empiricism to pre-dictability. Cryo-Letters 11, 93–110.

Franks, F., 1993. Solid aqueous solutions. Pure Appl. Chem.65, 2527–2537.

Franks, F., 1994. Long-term stabilization of biologicals. Bio-technology 12, 253–256.

Fransson, J., Florin-Robertsson, E., Axelsson, K., Nyhlen, C.,1996. Oxidation of human insulin-like growth factor I informulation studies: kinetics of methionine oxidation inaqueous solution and in solid state. Pharm. Res. 13, 1252–1257.

French, D.L., Arakawa, T., Niven, R.W., Li, T., 1995. FTIRinvestigation of hydration-induced conformational changesin spray-dried protein/trehalose powders. Pharm. Res. 12(Suppl.), S83.

Getlin, L.A., 1991. Kinetics of phase transitions in a frozensolution. Dev. Biol. Stand. 74, 93–104.

Gibson, T.D., 1996. Protein stabilization using additives basedon multiple electrostatic interactions. Dev. Biol. Stand. 87,207–217.

Gombotz, W.R., Pankey, S.C., Bouchard, L.S., Phan, D.H.,MacKenzie, A.P., 1996. Stability, characterization, formu-lation and delivery system development for transforminggrowth factor-beta1. In: Pearlman, R., Wang, Y.J. (Eds.),Formulation, Characterization, and Stability of ProteinDrugs. Plenum Press, New York, pp. 219–245.

Goormaghtigh, E., Cabiaux, V., Ruysschaert, J.-M., 1994.Determination of soluble and membrane protein structureby Fourier transform infrared spectroscopy II. Experimen-tal aspects, side chain structure, and H/D exchange. Sub-cell. Biochem. 23, 363–404.

Goto, Y., Fink, A.L., 1989. Conformational states of b-lacta-mase: molten-globule states at acidic and alkaline pH withhigh salt. Biochemistry 28, 945–952.

Graziano, G., Catanzano, F., Riccio, A., Barone, G., 1997. Areassessment of the molecular origin of cold denaturation.J. Biochem. 122, 395–401.

Greiff, D., 1992. Development of cycles for lyophilization.Dev. Biol. Stand. 74, 85–92.

Griebenow, K., Klibanov, A.M., 1995. Lyophilization-inducedreversible changes in the secondary structure of proteins.Proc. Natl. Acad. Sci. USA 92, 10969–10976.

Gu, L.C., Erds, E.A., Chiang, H.-S., Calderwood, T., Tsai, K.,Visor, G.C., Duffy, J., Hsu, W.-C., Foster, L.C., 1991.Stability of interleukin 1b(IL-1b) in aqueous solution: ana-lytical methods, kinetics, products, and solution formula-tion implications. Pharm. Res. 8, 485–490.

Hageman, M.J., 1988. The role of moisture in protein stability.Drug Dev. Ind. Pharm. 14, 2047–2069.

Hagen, S.J., Hofrichter, J., Eaton, W.A., 1995. Protein reac-tion kinetics in a room-temperature glass. Science 269,959–962.

Hagen, S.J., Hofrichter, J., Eaton, W.A., 1996. Germinaterebinding and conformational dynamics of myoglobin em-bedded in a glass at room temperature. J. Phys. Chem. 100,12008–12021.


Hancock, B.C., Zografi, G., 1994. The relationship betweenthe glass transition temperature and the water content ofamorphous pharmaceutical solids. Pharm. Res. 11, 471–477.

Hancock, B.C., Zografi, G., 1997. Characteristics and signifi-cance of the amorphous state in pharmaceutical systems. J.Pharm. Sci. 86, 1–12.

Hancock, B.C., Shamblin, S.L., Zografi, G., 1995. Molecularmobility of amorphous pharmaceutical solids below theirglass transition temperature. Pharm. Res. 12, 799–806.

Hancock, B.C., Christensen, K., Shamblin, S.L., 1998. Esti-mating the critical molecular mobility temperature (TK) ofamorphous pharmaceuticals. Pharm. Res. 15, 1649–1651.

Hartmann, U., Bunner, B., Korber, C.H., Rau, G., 1991.Where should be cooling rate be determined in an extendedfreezing sample. Cryobiology 28, 115–130.

Hatley, R.H., 1997. Glass fragility and the stability of pharma-ceutical preparations-excipient selection. Pharm. Dev.Technol. 2, 257–264.

Hatley, R.H.M., Franks, F., 1991. Applications of DSC in thedevelopment of improved freeze-drying processes for labilebiologicals. J. Therm. Anal. 37, 1905–1914.

Hatley, R.H.M., Franks, F., Brown, S., Sandhu, G., Gray,M., 1996. Stabilization of a pharmaceutical drug substanceby freeze-drying: a case study. Drug Stability 1, 2–12.

Hekman, C., Park, S., Teng, W.Y., Guzman, N.A., Rossi, T.,1995. Degradation of lyophilized and reconstitutedMacroscint. J. Pharm. Biomed. Anal. 13, 1249–1261.

Heimburg, T., Marsh, D., 1993. Investigation of secondaryand tertiary structural changes of cytochrome c in com-plexes with anionic lipids using amide hydrogen exchangemeasurements: an FTIR study. Biophys. J. 65, 2408–2417.

Heller, M.C., Carpenter, J.F., Randolph, T.W., 1997. Manipu-lation of lyophilization-induced phase separation: implica-tions for pharmaceutical proteins. Biotechnol. Prog. 13,590–596.

Heller, M.C., Carpenter, J.F., Randolph, T.W., 1999a. Proteinformulation and lyophilization cycle design: prevention ofdamage due to freeze-concentration induced phase separa-tion. Biotechnol. Bioeng. 63, 166–174.

Heller, M.C., Carpenter, J.F., Randolph, T.W., 1999b. Con-formational stability of lyophilized PEGylated proteins ina phase-separating system. J. Pharm. Sci. 88, 58–64.

Heller, M.C., Carpenter, J.F., Randolph, T.W., 1999c. Appli-cation of a thermodynamic model to the prediction ofphase separations in freeze-concentrated formulations forprotein lyophilization. Arch. Biochem. Biophys. 363, 191–201.

Her, L.-M., Nail, S.L., 1994. Measurement of glass transitiontemperatures of freeze-concentrated solutes by differentialscanning calorimetry. Pharm. Res. 11, 54–59.

Hora, M.S., Rana, R.K., Smith, F.W., 1992a. Lyophilizedformulations of recombinant tumor necrosis factor.Pharm. Res. 9, 33–36.

Hora, M.S., Rana, R.K., Wilcox, C.L., Katre, N.V., Hirtzer,P., Wolfe, S.N., Thomson, J.W., 1992b. Development of alyophilized formulation of interleukin-2. Dev Biol Stand.74, 295–303.

House, J.A., Mariner, J.C., 1996. Stabilization of rinderpestvaccine by modification of the lyophilization process. Dev.Biol. Stand. 87, 235–244.

Hsu, C.C., Ward, C.A., Pearlman, R., Nguyen, H.M., Yeung,D.A., Curley, J.G., 1991. Determining the optimum resid-ual moisture in lyophilized protein pharmaceuticals. Dev.Biol. Stand. 74, 255–271.

Hsu, C.C., Nguyen, H.M., Yeung, D.A., Brookes, D.A., Koe,G.S., Bewley, T.A., Pearlman, R., 1995. Surface denatura-tion at solid-void interface — a possible pathway by whichopalescent particulates form during the storage oflyophilized tissue-type plasminogen activator at high tem-peratures. Pharm. Res. 12, 69–77.

Hsu, C.C., Walsh, A.J., Nguyen, H.M., Overcashier, D.E.,Koning-Bastiaan, H., Bailey, R.C., Nail, S.L., 1996. De-sign and application of a low-temperature Peltier-coolingmicroscope stage. J. Pharm. Sci. 85, 70–74.

Izutsu, K., Yoshioka, S., Terao, T., 1993. Decreased protein-stabilizing effects of cryoprotectants due to crystallization.Pharm. Res. 10, 1232–1237.

Izutsu, K., Yoshioka, S., Kojima, S., 1994a. Physical stabilityand protein stability of freeze-dried cakes during storage atelevated temperatures. Pharm. Res. 11, 995–999.

Izutsu, K., Yoshioka, S., Terao, T., 1994b. Effect of mannitolcrystallinity on the stabilization of enzymes during freeze-drying. Chem. Pharm. Bull. (Tokyo) 42, 5–8.

Izutsu, K., Yoshioka, S., Kojima, S., 1995. Increased stabiliz-ing effects of amphiphilic excipients on freeze-drying oflactate dehydrogenase (LDH) by dispersion into sugarmatrices. Pharm. Res. 12, 838–843.

Izutsu, K., Yoshioka, S., Kojima, S., Randolph, T.W., Car-penter, J.F., 1996. Effects of sugars and polymers oncrystallization of poly(ethylene glycol) in frozen solutions:phase separation between incompatible polymers. Pharm.Res. 13, 1393–1400.

Jaenicke, R., 1990. Protein structure and function at lowtemperatures. Philos. Trans. R. Soc. Lond. B Biol. Sci.326, 535–551.

Jennings, T.A., 1999. Lyophilization: Introduction and BasicPrinciples. Interpharm Press, Colorado, USA.

Jennings, T.A., Duan, N., 1995. Calorimetric monitoring oflyophilization. J. Parent. Sci. Technol. 49, 272–282.

Jordan, G.M., Yoshioka, S., Terao, T., 1994. The aggregationof bovine serum albumin in solution and in the solid state.J. Pharm. Pharmacol. 46, 182–185.

Kasraian, K., Spitznagel, T.M., Juneau, J.A., Yim, K., 1998.Characterization of the sucrose/glycine/water system bydifferential scanning calorimetry and freeze-drying mi-croscopy. Pharm. Dev. Technol. 3, 233–239.

Katakam, M., Banga, A.K., 1995. Aggregation of insulin andits prevention by carbohydrate excipients. PDA J. Pharm.Sci. Technol. 49, 160–165.

Katakam, M., Robillard, P., Brown, J., Tolman, G., 1998.Effects of moisture and lyophilization cycle on aggregationof a monoclonal antibody product. Pharm. Sci. 1 (Suppl.),S543.


Kendrick, B., Chang, B.S., Dong, A., Kreilgard, L., Carpen-ter, J.F., 1995a. An infrared spectroscopic study on thestructural changes induced by cooling and freezing stresseson proteins. Pharm. Res. 12 (Suppl.), 84.

Kendrick, B., Chang, B.Y., Carpenter, J.F., 1995b. Detergentstabilization of proteins against surface and freezing denat-uration. Pharm. Res., 12 (Supple), S-85.

Kendrick, B.S., Dong, A., Allison, S.D., Manning, M.C.,Carpenter, J.F., 1996. Quantitation of the area of overlapbetween second-derivative amide I infrared spectra to de-termine the structural similarity of a protein in differentstates. J. Pharm. Sci. 85, 155–158.

Kenney, W.C., Watson, E., Bartley, T., Boone, T., Altrock,B.W., 1986. Parameters for the evaluation of IL-2 stability.Lymphokine Res. 5 (Suppl. 1), S23–S27.

Kerr, W.L., Lim, M.H., Reid, D.S., Chen, H., 1993. Chemicalreaction kinetics in relation to glass transition temperaturein frozen food polymer solutions. J. Sci. Food Agric. 61,51–56.

Kerwin, B.A., Heller, M.C., Levin, S.H., Randolph, T.W.,1998. Effects of Tween 80 and sucrose on acute short-termstability and long-term storage at −20°C of a recombi-nant hemoglobin. J. Pharm. Sci. 87, 1062–1068.

Kim, A.I., Akers, M.J., Nail, S.L., 1998. The physical state ofmannitol after freeze-drying: effect of mannitol concentra-tion, freezing rate, and a noncrystallizing cosolute. J.Pharm. Sci. 87, 931–935.

Koseki, T., Kitabatake, N., Doi, E., 1990. Freezing denatura-tion of ovalbumin at acid pH. J. Biochem. 107, 389–394.

Kreilgaard, L., Frokjaer, S., Flink, J.M., Randolph, T.W.,1998a. Carpenter JF Effects of additives on the stability ofrecombinant human factor XIII during freeze-drying andstorage in the dried solid. Arch. Biochem. Biophys. 360,121–134.

Kreilgaard, L., Jones, L.S., Randolph, T.W., Frokjaer, S.,Flink, J.M., Manning, M.C., Carpenter, J.F., 1998b. Effectof Tween 20 on freeze-thawing- and agitation-inducedaggregation of recombinant human factor XIII. J. Pharm.Sci. 87, 1597–1603.

Kreilgaard, L., Frokjaer, S., Flink, J.M., Randolph, T.W.,Carpenter, J.F., 1999. Effects of additives on the stabilityof Humicola lanuginosa lipase during freeze-drying andstorage in the dried solid. J. Pharm. Sci. 88, 281–290.

Kristjansson, M.M., Kinsella, J.E., 1991. Protein and enzymestability: structural, thermodynamic, and experimental as-pects. Adv. Food Nutr. Res. 35, 237–316.

Kuhlman, B., Yang, H.Y., Boice, J.A., Fairman, R., Raleigh,D.P., 1997. An exceptionally stable helix from the riboso-mal protein L9: Implications for protein folding and stabil-ity. J. Mol. Biol. 270, 640–647.

Lai, M.C., Topp, E.M., 1999. Solid-state chemical stability ofproteins and peptides. J. Pharm. Sci. 88, 489–500.

Lam, X.M., Costantino, H.R., Overcashier, D.E., Nguyen,T.H., Hsu, C.C., 1996. Replacing succinate with glyco-cholate buffer improves the stability of lyophilized inter-feron-g. Int. J. Pharm. 142, 85–95.

Leslie, S.B., Israeli, E., Lighthart, B., Crowe, J.H., Crowe,L.M., 1995. Trehalose and sucrose protect both mem-branes and proteins in intact bacteria during drying. Appl.Environ. Microbiol. 61, 3592–3597.

Levine, H., Slade, L., 1992. Another view of trehalose fordrying and stabilizing biological materials. Biopharm. 5(5), 36–40.

Li, S., Nguyen, T., Schoneich, C., Borchardt, R.T., 1995a.Aggregation and precipitation of human relaxin inducedby metal-catalyzed oxidation. Biochemistry 34, 5762–5772.

Li, S., Schoneich, C., Borchardt, R.T., 1995b. Chemical insta-bility of protein pharmaceuticals: mechanisms of oxidationand strategies for stabilization. Biotechnol. Bioeng. 48,491–500.

Li, S., Patapoff, T.W., Overcashier, D., Hsu, C., Nguyen,T.H., Borchardt, R.T., 1996. Effects of reducing sugars onthe chemical stability of human relaxin in the lyophilizedstate. J. Pharm. Sci. 85, 873–877.

Librizzi, F., Vitrano, E., Cordone, L., 1999. Dehydration andcrystallization of trehalose and sucrose glasses containingcarbonmonoxy-myoglobin. Biophys. J. 76, 2727–2734.

Lin, T.-Y., Timasheff, S.N., 1996. On the role of surfacetension in the stabilization of globular proteins. ProteinSci. 5, 372–381.

Lin, L.S., Kunitani, M.G., Hora, M.S., 1996. Interferon-b-1b(Betaseron®): a model for hydrophobic therapeuticproteins. In: Pearlman, R., Wang, Y.J. (Eds.), Formula-tion, Characterization, and Stability of Protein Drugs.Plenum Press, New York, pp. 275–301.

Liu, R., Orgel, L., 1997. Efficient oligomerization of negativelycharged Beta-amino acids at −20°C. J. Am. Chem. Soc.119, 9192.

Liu, W.R., Langer, R., Klibanov, A.M., 1990. Moisture-in-duced aggregation of lyophilized proteins in the solid state.Biotechnol. Bioeng. 37, 177–184.

Liu, J.L., Lu, K.V., Eris, T., Katta, V., Westcott, K.R., Narhi,L.O., Lu, H.S., 1998. In vitro methionine oxidation ofrecombinant human leptin. Pharm. Res. 15, 632–640.

Lu, Q., Zografi, G., 1997. Properties of citric acid at the glasstransition. J. Pharm. Sci. 86, 1374–1378.

Lueckel, B., Bodmer, D., Helk, B., Leuenberger, H., 1998a.Formulations of sugars with amino acids or mannitol-infl-uence of concentration ratio on the properties of thefreeze-concentrate and the lyophilizate. Pharm. Dev. Tech-nol. 3, 325–336.

Lueckel, B., Helk, B., Bodmer, D., Leuenberger, H., 1998b.Effects of formulation and process variables on the aggre-gation of freeze-dried interleukin-6 (IL-6) after lyophiliza-tion and on storage. Pharm. Dev. Technol. 3, 337–346.

Ma, X., Wang, D.Q., Bouffard, R., MacKenzie, A., 1998.Characterization of protein formulation for freeze-dryingcycle development. Pharm. Sci. 1 (Suppl.), S543.

Manning, M.C., Patel, K., Borchardt, R.T., 1989. Stability ofprotein pharmaceuticals. Pharm. Res. 6, 903–917.

Mattern, M., Winter, G., Rudolph, R., Kohnert, U., Lee, G.,1997. Formulation of proteins in vacuum-dried glasses. I.Improved vacuum drying using crystallizing amino acids.Eur. J. Pharm. Biopharm. 44, 177–185.


Mattern, M., Winter, G., Kohnert, U., Lee, G., 1999. Formu-lation of proteins in vacuum-dried glasses. II. Process andstorage stability in sugar-free amino acid systems. Pharm.Dev. Technol. 4, 199–208.

May, J.C., Wheeler, R.M., Grim, E., 1989. The gravimetricmethod for the determination of residual moisture infreeze-dried biological products. Cryobiology 26, 277–284.

Mazzobre, M.F., Buera, M.P., Chirife, J., 1997. Glass transi-tion and thermal stability of lactase in low-moisture amor-phous polymeric matrices. Biotechnol. Prog. 13, 195–199.

Meredith, P., Donald, A.M., Payne, R.S., 1996. Freeze-drying:in situ observations using cryoenviromental scanning elec-tron microscopy and differential scanning calorimetry. J.Pharm. Sci. 85, 631–637.

Miller, D.P., de Pablo, J.J., Corti, H., 1997. Thermophysicalproperties of trehalose and its concentrated aqueous solu-tions. Pharm. Res. 14, 578–590.

Moreira, T., Pendas, J., Gutierrez, A., Pomes, R., Duque, J.,Franks, F., 1998. Effect of sucrose and raffinose on physi-cal state and on lactate dehydrogenase activity of freeze-dried formulations. Cryo-Letters 19, 115–122.

Mozhaev, V.V., Martinek, K., 1984. Structure-stability rela-tionships in proteins: new approaches to stabilizing en-zymes. Enzyme Microb. Technol. 6, 50–59.

Nail, S.L., Johnson, W., 1992. Methodology for in-processdetermination of residual water in freeze-dried products.Dev. Biol. Stand. 74, 137–150.

Nagendra, H.G., Sukumar, N., Vijayan, M., 1998. Role ofwater in plasticity, stability, and action of proteins: thecrystal structures of lysozyme at very low levels of hydra-tion. Proteins: Struct. Func. Gen. 32, 229–240.

Naini, V., Byron, P.R., Phillips, E.M., 1998. Physicochemicalstability of crystalline sugars and their spray-dried forms:dependence upon relative humidity and suitability for usein powder inhalers. Drug Dev. Ind. Pharm. 24, 895–909.

Nakai, Y., Yoshioka, S., Aso, Y., Kojima, S., 1998. Solid-staterehydration-induced recovery of bilirubin oxidase activityin lyophilized formulations reduced during freeze-drying.Chem. Pharm. Bull. 46, 1031–1033.

Nayar, R., 1998. Stabilized albumin-free recombinant factorVIII having a low sugar content. US patent c5763401.

Nema, S., Avis, K.E., 1992. Freeze-thaw studies of a modelprotein, lactate dehydrogenase, in the presence of cryopro-tectants. J. Parenter. Sci. Technol. 47, 76–83.

Nguyen, T.H., Burnier, J., Meng, W., 1993. The kinetics ofrelaxin oxidation by hydrogen peroxide. Pharm. Res. 10,1563–1571.

O’Brien, J., 1996. Stability of trehalose, sucrose and glucose tononenzymatic browning in model systems. J. Food Sci. 61,679–682.

Oetjen, G.-W., 1999. Industrial freeze-drying for pharmaceuti-cal applications. In: Rey, L., May, J.C. (Eds.), Freeze-Dry-ing/Lyophilization of Pharmaceutical and BiologicalProducts, vol. 96. Marcel Dekker, New York, pp. 267–335.

Oksanen, C.A., Zografi, G., 1990. The relationship betweenthe glass transition temperature and water vapor absorp-tion by poly(vinylpyrrolidone). Pharm. Res. 7, 654–657.

Oksanen, C.A., Zografi, G., 1993. Molecular mobility in mix-tures of absorbed water and solid poly(vinylpyrrolidone).Pharm. Res. 10, 791–799.

O8 sterberg, T., Wadsten, T., 1999. Physical state of L-histidineafter freeze-drying and long-term storage. Eur. J. Pharm.Sci. 8, 301–308.

O8 sterberg, T., Fatouros, A., Mikaelsson, M., 1997. Develop-ment of freeze-dried albumin-free formulation of recombi-nant factor VIII SQ. Pharm. Res. 14, 892–898.

Overcashier, D.E., Brooks, D.A., Costantino, H.R., Hsu,C.C., 1997. Preparation of excipient-free recombinant hu-man tissue-type plasminogen activator by lyophilizationfrom ammonium bicarbonate solution: An investigation ofthe two-stage sublimation phenomenon. J. Pharm. Sci. 86,455–459.

Overcashier, D.E., Patapoff, T.W., Hsu, C.C., 1999.Lyophilization of protein formulations in vials: investiga-tion of the relationship between resistance to vapor flowduring primary drying and small-scale product collapse. J.Pharm. Sci. 88, 688–695.

Paulsen, H., Pflughaupt, K.-W., 1980. Glycosylamines. In:Pigman, W., Horton, D., Wander, J.D. (Eds.), The Carbo-hydrates Chemistry and Biochemistry. Academic Press,New York, pp. 881–921.

Pearson, D.S., Smith, G., 1998. Dielectric analysis as a tool forinvestigating the lyophilization of proteins. Pharm. Sci.Technol. Today 1, 108–117.

Perlman, R., Nguyen, T., 1992. Pharmaceutics of proteindrugs. J. Pharm. Pharmcol. 44 (Suppl. 1), 178–185.

Physicians Desk Reference, 1999. Medical Economics Com-pany Incorporation, 53rd edn. Montvale, NJ.

Pikal, M.J., 1990a. Freeze-drying of proteins. Part I: processdesign. Biopharm. 3 (8), 18–27.

Pikal, M.J., 1990b. Freeze-drying of proteins. Part II: formula-tion selection. Biopharm. 3 (9), 26–30.

Pikal, M.J., 1999. Mechanisms of protein stabilization duringfreeze-drying and storage: the relative importance of ther-modynamic stabilization and glassy state relaxation dy-namics. In: Rey, L., May, J.C. (Eds.),Freeze-Drying/Lyophilization of Pharmaceutical and Bio-logical Products, vol. 96. Marcel Dekker, New York, pp.161–198.

Pikal, M.J., Rigsbee, D.R., 1997. The stability of insulin incrystalline and amorphous solids: observation of greaterstability for the amorphous form. Pharm. Res. 14, 1379–1387.

Pikal, M.J., Shah, S., 1997. Intravial distribution of moistureduring the secondary drying stage of freeze drying. PDA J.Pharm. Sci. Technol. 51, 17–24.

Pikal, K.A., Carpenter, J.F., 1998. pH changes during freezingin sodium and potassium phosphate buffer systems in thepresence of glycine: effect on protein stability. Pharm. Sci.1 (Suppl.), 544.

Pikal, M.J., Dellerman, K.M., Roy, M.L., Riggin, R.M., 1991.The effects of formulation variables on the stability offreeze-dried human growth hormone. Pharm. Res. 8, 427–436.


Pikal, M.J., Dellerman, K., Roy, M.L., 1992. Formulation andstability of freeze-dried proteins: effects of moisture andoxygen on the stability of freeze-dried formulations ofhuman growth hormone. Dev. Biol. Stand. 74, 21–37.

Prestrelski, S.J., Arakawa, T., Carpenter, J.F., 1993a. Separa-tion of freezing and drying-induced denaturation oflyophilized proteins using stress-specific stabilization. II.Structural studies using infrared spectroscopy. Arch.Biochem. Biophys. 303, 465–473.

Prestrelski, S.J., Tedeschi, N., Arakawa, T., Carpenter, J.F.,1993b. Dehydration-induced conformational transitions inproteins and their inhibition by stabilizers. Biophy. J. 65,661–671.

Prestrelski, S.J., Pikal, K.A., Arakawa, T., 1995. Optimizationof lyophilization conditions for recombinant human inter-leukin-2 by dried-state conformational analysis usingFourier-transform infrared spectroscopy. Pharm. Res. 12,1250–1259.

Ramos, A., Raven, N.D.H., Sharp, R.J., Bartolucci, S.,Roswsi, M., Cannio, R., Lebbink, J., vanderOost, J., de-Vos, W.M., Santos, H., 1997. Stabilization of enzymesagainst thermal stress and freeze-drying by mannosylglyc-erate. Appl. Environ. Microbiol. 63, 4020–4025.

Ravve, A., 1967. Physical properties of macromolecules.Chapter 2 in Organic Chemistry of Macromolecules,@Marcel Dekker, New York. pp. 17-36.

Remmele, R.L. Jr, Stushnoff, C., Carpenter, J.F., 1997. Real-time in situ monitoring of lysozyme during lyophilizationusing infrared spectroscopy: dehydration stress in the pres-ence of sucrose. Pharm. Res. 14, 1548–1555.

Ressing, M.E., Jiskoot, W., Talsma, H., van Ingen, C.W.,Beuvery, E.C., Crommelin, D.J., 1992. The influence ofsucrose, dextran, and hydroxypropyl-beta-cyclodextrin aslyoprotectants for a freeze-dried mouse IgG2a monoclonalantibody (MN12). Pharm. Res. 9, 266–270.

Rey, L., 1999. Glimpses into the realm of freeze-drying: classi-cal issues and new ventures. In: Rey, L., May, J.C. (Eds.),Freeze-Drying/Lyophilization of Pharmaceutical and Bio-logical Products, vol. 96. Marcel Dekker, New York, pp.1–30.

Reyes, F.G.R., Poocharoen, B., Wrolstad, R.E., 1982. Mail-lard browning reaction of sugar–glycine model systems:changes in sugar concentration, color and appearance. J.Food Sci. 47, 1376–1377.

Rossi, S., Buera, M.P., Moreno, S., Chirife, J., 1997. Stabiliza-tion of the restriction enzyme EcoRI dried with trehaloseand other selected glass-forming solutes. Biotechnol. Prog.13, 609–616.

Roy, M.L., Pikal, M.J., Rickard, E.C., Maloney, A.M., 1992.The effects of formulation and moisture on the stability ofa freeze-dried monoclonal antibody-vinca conjugate: a testof the WLF glass transition theory. Dev. Biol. Stand. 74,323–339.

Rupley, J.A., Careri, G., 1991. Protein hydration and func-tion. Adv. Protein Chem. 41, 37–173.

Roos, Y., 1993. Melting and glass transitions of low molecularweight carbohydrates. Carbohydr. Res. 238, 39–48.

Roos, Y., Karel, M., 1991. Phase transitions of mixtures ofamorphous polysaccharides and sugars. Biotechnol. Prog.7, 49–53.

Roser, B., 1991. Trehalose drying: a novel replacement forfreeze-drying. Biopharm. 4 (9), 47–53.

Ruddon, R.W., Bedows, E., 1997. Assisted protein folding. J.Biol. Chem. 272, 3125.

Sampedro, J.G., Guerra, G., Pardo, J.P., Uribe, S., 1998.Trehalose-mediated protection of the plasma membraneH+ -ATPase from Kluyveromyces lactis during freeze-dry-ing and rehydration. Cryobiology 37, 131–138.

Sarciaux, J.M., Hageman, M.J., 1997. Effects of bovine soma-totropin (rbSt) concentration at different moisture levelson the physical stability of sucrose in freeze-dried rbSt/su-crose mixtures. J. Pharm. Sci. 86, 365–371.

Sarciaux, J.-M., Mansour, S., Hageman, M.J., Nail, S.L.,1998. Effect of species, processing conditions and phos-phate buffer composition on IgG aggregation duringlyophilization. Pharm. Sci. 1 (Suppl.), 545.

Sasaoki, K., Hiroshima, T., Kusumoto, S., Nishu, K., 1992.Deamidation at asparagine-88 in recombinant human in-terleukin 2. Chem. Pharm. Bull. (Tokyo) 40, 976–980.

Sartor, G., Mayer, E., Johari, G.P., 1994. Calorimetric studiesof the kinetic unfreezing of molecular motions in hydratedlysozyme, hemoglobin, and myoglobin. Biophys. J. 66,249–258.

Sastry, G.M., Agmon, N., 1997. Trehalose prevents myoglobincollapse and preserves its internal mobility. Biochemistry36, 7097–7108.

Savage, M., Torres, J., Franks, L., Masecar, B., Hotta, J.,1998. Determination of adequate moisture content forefficient dry-heat viral inactivation in lyophilized factorVIII by loss on drying and by near infrared spectroscopy.Biologicals 26, 119–124.

Schebor, C., Buera, M.P., Chirife, J., 1996. Glassy state inrelation to the thermal inactivation of enzyme invertase inamorphous dried matrices of trehalose, maltodextrin andPVP. J. Food Eng. 30, 269–282.

Schebor, C., Burin, L., Buera, M.P., Aguilera, J.M., Chirife,J., 1997. Glassy state and thermal inactivation of invertaseand lactase in dried amorphous matrices. Biotechnol. Prog.13, 857–863.

Schmitt, E.A., Law, D., Zhang, G.G.Z., 1999. Nucleation andcrystallization kinetics of hydrated amorphous lactoseabove glass transition temperature. J. Pharm. Sci. 88,291–296.

Schwendeman, S.P., Costantino, H.R., Gupta, R.K., Siber,G.R., Klibanov, A.M., Langer, R., 1995. Stabilization oftetanus and diphtheria toxoids against moisture-inducedaggregation. Proc. Natl. Acad. Sci. USA 92, 11234–11238.

Schwendeman, S.P., Costantino, H.R., Gupta, R.K., Tobio,M., Chang, A.C., Alonso, M.J., Siber, G.R., Langer, R.,1996. Strategies for stabilizing tetanus toxoid towards thedevelopment of a single-dose tetanus vaccine. Dev. Biol.Stand. 87, 293–306.

Separovic, F., Lam, Y.H., Ke, X., Chan, H.K., 1998. Asolid-state NMR study of protein hydration and stability.Pharm. Res. 15, 1816–1821.


Shalaev, E.Y., Zografi, G., 1996. How does residual wateraffect the solid-state degradation of drugs in the amor-phous state? J. Pharm. Sci. 85, 1137–1141.

Shamblin, S.L., Zografi, G., 1999. The effects of absorbedwater on the properties of amorphous mixtures containingsucrose. Pharm. Res. 16, 1119–1124.

Shamblin, S.L., Huang, E.Y., Zografi, G., 1996. The effects ofcolyophilized polymeric additives on the glass transitiontemperature and crystallization of amorphous sucrose. J.Therm. Anal. 47, 1567–1579.

Shamblin, S.L., Hancock, B.C., Zografi, G., 1998. Watervapor sorption by peptides, proteins and their formula-tions. Eur. J. Pharm. Biopharm. 45, 239–247.

Sheehan, P., Liapis, A.I., 1998. Modeling of the primary andsecondary drying stages of the freeze drying of pharmaceu-tical products in vials: numerical results obtained from thesolution of a dynamic and spatially multi-dimensionallyophilization model for different operational policies. Bio-technol. Bioeng. 60, 712–728.

Shikama, K., Yamazaki, I., 1961. Denaturation of catalase byfreezing and thawing. Nature 190, 83–84.

Shortle, D., 1996. The denatured state (the other half of thefolding equation) and its role in protein stability. FASEBJ. 10, 27–34.

Skrabanja, A.T., de Meere, A.L., de Ruiter, R.A., van denOetelaar, P.J., 1994. Lyophilization of biotechnology prod-ucts. PDA J. Pharm. Sci. Technol. 48, 311–317.

Slade, L., Levine, H., Finley, J.W., 1989. Protein-water inter-actions: water as a plasticizer of gluten and other proteinpolymers. In: Philips, R.D., Finley, J.W. (Eds.), ProteinQuality and the Effects of Processing. Marcel Dekker, NewYork, pp. 9–124.

Sola-Penna, M., Meyer-Fernandes, J.R., 1998. Stabilizationagainst thermal inactivation promoted by sugars on en-zyme structure and function: why is trehalose more effec-tive than other sugars? Arch. Biochem. Biophys. 360,10–14.

Strambini, G.B., Gabellieri, E., 1996. Proteins in frozen solu-tions: evidence of ice-induced partial unfolding. Biophys. J.70, 971–976.

Strickley, R.G., Anderson, B.D., 1996. Solid-state stability ofhuman insulin I. Mechanism and the effect of water on thekinetics of degradation in lyophiles from pH 2–5 solutions.Pharm. Res. 13, 1142–1153.

Strickley, R.G., Anderson, B.D., 1997. Solid-state stability ofhuman insulin I. Effect of water on reactive intermediatepartitioning in lyophiles from pH 2–5 solutions: stabiliza-tion against covalent dimer formation. J. Pharm. Sci. 86,645–653.

Sun, W.Q., 1997. Temperature and viscosity for structuralcollapse and crystallization of amorphous carbohydratesolutions. Cryo-Letters 18, 99–106.

Sun, W.Q., Davidson, P., 1998. Protein inactivation in amor-phous sucrose and trehalose matrices: effects of phaseseparation and crystallization. Biochim. Biophys Acta1425, 235–244.

Sun, W.Q., Davidson, P., Chan, H.S.O., 1998. Protein stabilityin the amorphous carbohydrate matrix: relevance to anhy-drobiosis. Biochim. Biophy. Acta 1425, 245–254.

Susi, H., Byler, D.M., 1986. Resolution-enhanced Fouriertransform infrared spectroscopy of enzymes. Methods En-zymol. 130, 290–311.

Tanaka, K., Takeda, T., Miyajima, K., 1991. Cryoprotectiveeffect of saccharides on denaturation of catalase by freeze-drying. Chem. Pharm. Bull. 39, 1091–1094.

Tarelli, E., Mire-Sluis, A., Tivnann, H.A., Bolgiano, B.,Crane, D.T., Gee, C., Lemercinier, X., Athayde, M.L.,Sutcliffe, N., Corran, P.H., Rafferty, B., 1998. Recombi-nant human albumin as a stabilizer for biological materialsand for the preparation of international reference reagents.Biologicals 26, 331–346.

Taylor, L.S., Zografi, G., 1998a. The quantitative analysis ofcrystallinity using FT-Raman spectroscopy. Pharm. Res.15, 755–761.

Taylor, L.S., Zografi, G., 1998b. Sugar-polymer hydrogenbond interactions in lyophilized amorphous mixtures. J.Pharm. Sci. 87, 1615–1621.

te Booy, M.P., de Ruiter, R.A., de Meere, A.L., 1992. Evalua-tion of the physical stability of freeze-dried sucrose-con-taining formulations by differential scanning calorimetry.Pharm. Res. 9, 109–114.

Timasheff, S.N., 1993. The control of protein stability andassociation by weak interactions with water: how do sol-vents affect these processes? Annu. Rev. Biophy. Biomol.Struct. 22, 67–97.

Timasheff, S.N., 1998. Control of protein stability and reac-tions by weakly interacting cosolvents: the simplicity of thecomplicated. Adv. Protein Chem. 51, 355–432.

Towns, J.K., 1995. Moisture content in proteins: its effects andmeasurement. J. Chromatogr. A705, 115–127.

Townsend, M.W., DeLuca, P.P., 1990. Stability of ribonucle-ase A in solution and the freeze-dried state. J. Pharm. Sci.79, 1083–1086.

Tsai, A.M., van Zanten, J.H., Betenbaugh, M.J., 1998. I.Study of protein aggregation due to heat denaturation: astructural approach using circular dichroism spectroscopy,nuclear magnetic resonance, and static light scattering.Biotechnol. Bioeng. 59, 273–280.

Tzannis, S.T., Prestrelski, S.J., 1999a. Activity-stability consid-erations of trypsinogen during spray drying: effects ofsucrose. J. Pharm. Sci. 88, 351–359.

Tzannis, S.T., Prestrelski, S.J., 1999b. Moisture effects onprotein-excipient interactions in spray-dried powders. Na-ture of destabilizing effects of sucrose. J. Pharm. Sci. 88,360–370.

Uritani, M., Takai, M., Yoshinaga, K., 1995. Protective effectof disaccharides on restriction endonucleases during dryingunder vacuum. J. Biochem. (Tokyo) 117, 774–779.

Vemuri, S., Yu, C.D., Roosdorp, N., 1994. Effect of cryopro-tectants on freezing, lyophilization, and storage oflyophilized recombinant alpha 1-antitrypsin formulations.PDA J. Pharm. Sci. Technol. 48, 241–246.


Volkin, D.B., Klibanov, A.M., 1989. Minimizing protein in-activation. In: Creighton, T.E. (Ed.), Protein Function APractical Approach. Information Press, Oxford, UK, pp.1–24.

Volkin, D.B., Middaugh, C.R., 1996. The characterization,stabilization, and formulation of acidic fibroblast growthfactor. In: Pearlman, R., Wang, Y.J. (Eds.), Formula-tion, Characterization, and Stability of Protein Drugs.Plenum Press, New York, pp. 181–217.

Vrkljan, M., Foster, T.M., Powers, M.E., Henkin, J., Porter,W.R., Staack, H., Carpenter, J.F., Manning, M.C., 1994.Thermal stability of low molecular weight urokinase dur-ing heat treatment. Effect of polymeric additives. Pharm.Res. 11 (2), 1004–1008.

Wang, W., 1996. Oral protein drug delivery. J. Drug Target4, 195–232.

Wang, W., 1999. Instability, stabilization, and formulationof liquid protein pharmaceuticals. Int. J. Pharm. 185,129–188.

Wang, Y.J., Chen, H., Busch, M., Baldwin, P.A., 1997.Headspace analysis for parenteral products: oxygen per-meation and integrity test. Pharm. Technol. 21, 108–122.

Ward, K.R., Adams, G.D., Alpar, H.O., Irwin, W.J., 1999.Protection of the enzyme L-asparaginase during lyophili-sation — a molecular modelling approach to predict re-quired level of lyoprotectant. Int. J. Pharm. 187,153–162.

Webb, S., Randolph, T., Carpenter, J., Cleland, J., 1998. Aninvestigation into reconstitution of lyophilized recombi-nant human interferon-gamma. Pharm. Sci. 1 (Suppl.),542.

Willemer, H., 1992. Measurements of temperatures, ice evap-oration rates and residual moisture contents in freeze-drying. Dev. Biol. Stand. 74, 123–134.

Willemer, H., 1999. Experimental freeze-drying: proceduresand equipment. In: Rey, L., May, J.C. (Eds.), Freeze-Drying/Lyophilization of Pharmaceutical and BiologicalProducts, vol. 96. Marcel Dekker, New York, pp. 79–121.

Wisniewski, R., 1998. Developing large-scale cryopreserva-tion systems for biopharmaceutical products. Biopharm.11, 50–60.

Wolkers, W.F., Oldenhof, H., Alberda, M., Hoekstra, F.A.,1998a. A Fourier transform infrared microspectroscopystudy of sugar glasses: application to anhydrobiotichigher plant cells. Biochim. Biophys. Acta 1379, 83–96.

Wolkers, W.F., van Kilsdonk, M.G., Hoekstra, F.A., 1998b.Dehydration-induced conformational changes of poly-L-lysine as influenced by drying rate and carbohydrates.Biochim. Biophys. Acta 1425, 127–136.

Wong, D., Parascrampuria, J., 1997. Pharmaceutical excipi-ents for the stabilization of proteins. Pharm. Technol. 21,34–50.

Wu, S., Leung, D., Vemuri, S., Shah, D., Yang, B., Wang,J., 1998. Degradation mechanism of lyophilized basicfibroblast growth factor (BFGF) protein in sugar formu-lation. Pharm. Sci. 1 (Suppl.), S-539.

Yang, T.-H., Dong, A., Meyer, J., Johnson, O.L., Cleland,J.L., Carpenter, J.F., 1999. Use of infrared spectroscopyto assess secondary structure of human growth hormonewithin biodegradable microspheres. J. Pharm. Sci. 88,161–165.

Yeo, S.-D., Debenedetti, P.G., Patro, S.Y., Przybycien,T.M., 1994. Secondary structure characterization of mi-croparticulate insulin powders. J. Pharm. Sci. 83, 1651–1656.

Yoshioka, S., Aso, Y., Izutsu, K., Terao, T., 1993. Aggre-gates formed during storage of b-galactosidase in solu-tion and in the freeze-dried state. Pharm. Res. 10,103–108.

Yoshioka, S., Aso, Y., Izutsu, K.-I., Terao, T., 1994. Appli-cation of accelerated testing for shelf-life prediction ofcommercial protein preparations. J. Pharm. Sci. 83, 454–456.

Yoshioka, S., Aso, Y., Kojima, S., 1997. Dependence of themolecular mobility and protein stability of freeze-driedg-globulin formulations on the molecular weight of dex-tran. Pharm. Res. 14, 736–741.

Yoshioka, S., Aso, Y., Nakai, Y., Kojima, S., 1998. Effectof high molecular mobility of poly(vinyl alcohol) onprotein stability of lyophilized g-globulin formulations. J.Pharm. Sci. 87, 147–151.

Yoshioka, S., Aso, Y., Kojima, S., 1999. The effect of excip-ients on the molecular mobility of lyophilized formula-tions, as measured by glass transition temperature andNMR relaxation-based critical mobility temperature.Pharm. Res. 16, 135–140.

Yu, L., Milton, N., Groleau, E.G., Mishra, D.S., Vansickle,R.E., 1999. Existence of a mannitol hydrate duringfreeze-drying and practical implications. J. Pharm. Sci.88, 196–198.

Zhang, M.Z., Wen, J., Arakawa, T., Prestrelski, S.J., 1995.A new strategy for enhancing the stability of lyophilizedprotein: the effect of the reconstitution medium on ker-atinocyte growth factor. Pharm. Res. 12, 1447–1452.

Zhang, M.Z., Pikal, K., Nguyen, T., Arakawa, T., Prestrel-ski, S.J., 1996. The effect of the reconstitution mediumon aggregation of lyophilized recombinant interleukin-2and ribonuclease A. Pharm. Res. 13, 643–646.

.

wei wang

Documents