protein quantity at the air-solid interface publication

RESEARCH ARTICLE – Pharmaceutical Biotechnology

Protein Quantity on the Air–Solid Interface DeterminesDegradation Rates of Human Growth Hormone in LyophilizedSamples

YEMIN XU,1 PAWEL GROBELNY,2 ALEXANDER VON ALLMEN,3 KORBEN KNUDSON,3 MICHAEL PIKAL,2

JOHN F. CARPENTER,4 THEODORE W. RANDOLPH3

1Center for Pharmaceutical Biotechnology, Department of Biochemistry, University of Colorado, Boulder, Colorado 803092Department of Pharmaceutical Sciences, University of Connecticut, Storrs, Connecticut 062693Center for Pharmaceutical Biotechnology, Department of Chemical and Biological Engineering, University of Colorado, Boulder,Colorado 803094Center for Pharmaceutical Biotechnology, Department of Pharmaceutical Sciences, University of Colorado Denver, Anschutz MedicalCampus, Aurora, Colorado 80045

Received 22 January 2014; revised 11 February 2014; accepted 17 February 2014

Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23926

ABSTRACT: Recombinant human growth hormone (rhGH) was lyophilized with various glass-forming stabilizers, employing cycles thatincorporated various freezing and annealing procedures to manipulate glass formation kinetics, associated relaxation processes, and glass-specific surface areas (SSAs). The secondary structure in the cake was monitored by infrared and in reconstituted samples by circulardichroism. The rhGH concentrations on the surface of lyophilized powders were determined from electron spectroscopy for chemicalanalysis. Glass transition temperature (Tg), SSAs, and water contents were determined immediately after lyophilization. Lyophilized sampleswere incubated at 323 K for 16 weeks, and the resulting extents of rhGH aggregation, oxidation, and deamidation were determined afterrehydration. Water contents and Tg were independent of lyophilization process parameters. Compared with samples lyophilized after rapidfreezing, rhGH in samples that had been annealed in frozen solids prior to drying, or annealed in glassy solids after secondary dryingretained more native-like protein secondary structure, had a smaller fraction of the protein on the surface of the cake, and exhibited lowerlevels of degradation during incubation. A simple kinetic model suggested that the differences in the extent of rhGH degradation duringstorage in the dried state between different formulations and processing methods could largely be ascribed to the associated levels of rhGHat the solid–air interface after lyophilization. C© 2014 Wiley Periodicals, Inc. and the American Pharmacists Association J Pharm SciKeywords: protein structure; protein formulation; lyophilization; stability; mobility; specific surface area; annealing; surface degradation;growth hormone

INTRODUCTION

Lyophilization is widely accepted as an effective method to im-prove long-term stability of pharmaceuticals, especially thera-peutic protein products.1 In glassy lyophilized solids, both phys-ical and chemical degradation processes are greatly hindered.1

However, storage of proteins in glassy solid formulations doesnot always guarantee a desired shelf life.2,3 For decades, effortshave been made to choose appropriate formulations and designrobust lyophilization cycles to yield stable protein products.1 Ithas been widely documented that disaccharides (sucrose andtrehalose) are effective stabilizers during lyophilization andstorage, and this stabilizing effect has been ascribed to ther-modynamic and/or kinetic stabilization mechanisms.4–6 In ad-dition, the lyophilization cycle applied to a given formulationmay determine not only the morphology of the cake and phys-ical properties of the glass, but also the stability of the pro-tein during storage in the dried formulation.1,4,6–9 For exam-ple, the stability of methionyl human growth hormone in dried

Correspondence to: Theodore W. Randolph (Telephone: +303-492-4776;Fax: +303-492-8592; E-mail: [email protected])

Journal of Pharmaceutical SciencesC© 2014 Wiley Periodicals, Inc. and the American Pharmacists Association

solid formulations was shown to depend not only on the typeof stabilizer included in the formulation, but also on the dryingmethod that was used.9 Despite some insights into mechanismsby which excipients provide stability to proteins, developing aformulation that provides adequate protection against proteindamage is still a semiempirical exercise, as is the design of alyophilization cycle that provides optimal protein stability fora given formulation.

A conventional lyophilization cycle consists of freezing, pri-mary drying, and secondary drying steps. The freezing step isof paramount importance.10,11 During freezing, most of the wa-ter present in the original liquid (about 80%10) is crystallizedinto essentially pure ice, which results in a freeze-concentratedsolution for the remaining formulation. The pH of the freeze-concentrated liquid may undergo a shift because of the prefer-ential precipitation of buffer components, potentially contribut-ing to protein denaturation.12,13 In addition, the rates of icenucleation and crystal growth have large impacts on the icemorphology and ice–liquid interfacial area, and consequentlyon the final solid products’ specific solid–air interfacial areaformed after drying.7,10,14,15

Several strategies10,16 have been developed to manipulatethe initial stages of the lyophilization cycle, such as shelf-rampcooling, annealing, controlled ice nucleation, and fast freezing

Xu et al., JOURNAL OF PHARMACEUTICAL SCIENCES 1

2 RESEARCH ARTICLE – Pharmaceutical Biotechnology

by liquid N2. Shelf-ramp cooling is a standard cooling methodin commercial freeze-drying, during which shelf temperatureof the lyophilizer is decreased in a roughly linear fashion. Inthis case, the maximum cooling rate is limited by the cool-ing capacity of lyophilizer.17,18 Shelf-ramp cooling typically re-sults in a high degree of supercooling prior to the initiation offreezing, followed by a period characterized by rapid nucleationof a large number of ice crystals.14 Consequently, many smallice crystals are formed. During the primary drying portion ofthe lyophilization process, ice crystals are removed by sublima-tion, and the interface between the glass and the voids left be-hind contribute the specific surface area (SSA) of the resultingglassy solid.15,19,20 The large number of relatively small ice crys-tals formed after shelf-ramp freezing in turn yields lyophilizedcakes with large surface areas.10

Annealing refers to an additional step that may be addedafter freezing, during which the sample temperature is main-tained between the ice melting temperature and the glass tran-sition temperature of the maximally freeze-concentrated solu-tion, Tg

′11,19 (or the eutectic melting temperature of crystallineexcipients, if that temperature is greater than Tg

′). Becauseof the both enhanced mobility of water and the contributionsof surface energy to the elevated chemical potential of waterin smaller crystals, water is transported from small ice crys-tals and redeposits onto large ice crystals.10 As a result of thisOstwald ripening, larger ice crystals are generated during an-nealing, which in turn results in lyophilized cakes with reducedSSAs.11 For the purposes of this report, this type of annealingprocess will be termed “predrying annealing.”

We note that Ostwald ripening also occurs in lyophilizationcycles that use shelf-ramp cooling without an annealing step,but to a lesser degree, because the length of time available forripening (the time during which ice crystals are present andthe temperature is above Tg

′) is much shorter in these cycles.To reduce the extent of Ostwald ripening even further, fastfreezing methods may be used to limit the time that samplesspend at temperatures between the ice melting temperatureand Tg

′. One method of fast freezing is to immerse samplescontained in vials into liquid N2 (N2 immersion). Fast freezingalso may be achieved by spraying liquid droplets of sampledirectly into liquid N2 (N2-droplet freezing).13,21–24 In the N2

immersion procedure, the relatively low thermal conductivity ofthe glass vials limits heat transfer, making the effective coolingrate slower than that which can be achieved using the N2-droplet-freezing method.10 Compared with the standard shelf-ramp cooling, both of these two fast freezing methods result insmaller ice crystals and larger glassy matrix-SSAs.10

Another type of annealing (herein termed “postdrying an-nealing”) may be implemented by briefly incubating dry, glassysamples at a high (but sub-Tg) temperature at the end of thesecondary drying step of the lyophilization cycle. Wang et al.25

reported that postdrying annealing could enhance protein sta-bility. The stability increase was presumably a result of relax-ation processes that lead to slower motions in the glassy state.25

The current study examined formulations of recombinanthuman growth hormone (rhGH) in the presence of three glass-forming excipients: sucrose, trehalose, and hydroxyethyl starch(HES). These formulations were lyophilized using five differ-ent methods, which yielded glassy solids with different SSAs,surface protein contents, glassy state mobilities, and degreesof retention of native secondary structure. Because we antici-pate that protein molecules located on the surface of lyophilized

glassy solids will have significantly faster degradation rates, wehypothesize that the extent of rhGH degradation during stor-age in various dried solid formulations prepared by differentprocessing methods can largely be ascribed to the resultinglevels of rhGH found at the solid-air interface after lyophiliza-tion.

MATERIALS AND METHODS

Materials

Recombinant human growth hormone was expressed in Es-cherichia coli and purified as described previously.8,26 HES (Vi-astarch) was purchased from Fresenius (Graz, Austria), and su-crose and trehalose were purchased from Mallinckrodt Baker(Phillipsburg, New Jersey). All other chemicals were purchasedas reagent grade or higher. Lyophilization glass vials (5 mL,product number 68000318) and butyl rubber stoppers (productnumber 19560042) were purchased from West PharmaceuticalServices (Linville, Pennsylvania).

Methods

Formulation and Lyophilization Cycle Design

Recombinant human growth hormone was formulated at a con-centration of 1 mg/mL in one of the three formulations. In addi-tion to rhGH, each formulation contained 2 mM sodium phos-phate at a pH of 7.4, as well as 5% (w/v) of HES, trehalose, orsucrose. Lyophilization was performed using a FTS LyostarI system. An aliquot (1 mL) of each rhGH formulation waspipetted into vials, and lyophilized with one of five dif-ferent lyophilization cycles, denoted as standard lyophiliza-tion, predrying annealing lyophilization, postdrying annealinglyophilization, N2 immersion lyophilization, and N2-droplet-freezing lyophilization.

In the standard lyophilization cycle, sample vials wereloaded onto the shelf, which was at room temperature. Theshelf temperature was reduced to 10◦C, and samples were equi-librated at this temperature for 1 h. Shelf temperature wasthen decreased to −5◦C at 1◦C/min, kept at −5◦C for 20 min,and then decreased to −45◦C at 1.3◦C/min. Samples were keptfrozen at −45◦C for 400 min. Primary drying was then initiatedand performed at a shelf temperature of −20◦C and a chamberpressure of 70 mTorr for 1400 min. Secondary drying was thenstarted by increasing the shelf temperature to 33◦C at a rateof 0.3◦C/min. Samples were held at 33◦C and 70 mTorr for 4 h.Finally, vials were sealed in the chamber under dry nitrogen.

For the predrying annealing lyophilization cycle, an addi-tional annealing step was added to the standard cycle. Aftersamples were kept frozen at −45◦C for 400 min, shelf tempera-ture was increased to −5◦C over 30 min. Then, shelf tempera-ture was kept at −5◦C for 6 h before cooling to −45◦C. The shelftemperature was kept at −45◦C for 6 h, and then the primaryand secondary drying steps followed the same protocol as in thestandard lyophilization cycle.

Postdrying annealing was performed using the same protocolas standard lyophilization cycle, except that after the standardsecondary drying step, shelf temperature was increased up to50◦C at a rate of 0.3◦C/min, and held at 50◦C for 6 h beforeending the cycle.

Liquid N2 immersion lyophilization was carried out by im-mersing the glass vial containing 1 mL formulation into a liquid

Xu et al., JOURNAL OF PHARMACEUTICAL SCIENCES DOI 10.1002/jps.23926

RESEARCH ARTICLE – Pharmaceutical Biotechnology 3

N2 bath for 2 min and then putting the vials onto the lyophilizershelf, which was precooled to −45◦C, for 400 min. The rest ofprimary and secondary drying steps were the same as standardlyophilization cycle.

In the liquid N2-droplet-freezing lyophilization cycle, sam-ples were slowly pipetted into the glass vials, which were filledwith liquid N2. The sample vials were quickly moved onto thelyophilizer shelf, which was precooled shelf at −45◦C, and heldat this temperature for 400 min. The rest of the primary andsecondary drying cycle followed the same protocol as in thestandard lyophilization cycle.

Measurement of Residual Water Content

Residual water contents of the lyophilized samples were ana-lyzed using the Karl Fischer method.27 Triplicate samples wereprepared in a dry nitrogen-purged box and measured using aMettler DL37 KF coulometer (Hightstown, New Jersey), as de-scribed previously.8

Glass Transition Temperature Measurement by DifferentialScanning Calorimetry

The glass transition temperatures of the maximally freeze-concentrated formulation (Tg

′) and the lyophilized formulations(Tg) were measured with a PerkinElmer Diamond differentialscanning calorimetry (DSC) instrument. For Tg

′ measurement,aqueous solutions (20 :L) in aluminum pans were cooled fromroom temperature to −60◦C at 10◦C/min. After equilibration at−60◦C/min for 5 min, samples were heated to −5◦C at 5◦C/min,and kept at −5◦C for 30 min before samples were recooled to−60◦C again at 10◦C/min. The second heating scan was up to10◦C at 5◦C/min. In order to eliminate any thermal history ef-fects, Tg

′ was determined from the onset of thermal transitionmeasured during the second heating scan. Tg measurement ofthe lyophilization products followed the same protocol as de-scribed previously.8 At least triplicate samples were used todetermine the Tg and Tg

′.

Protein Secondary Structure by Infrared Spectroscopy

In a dry box, lyophilized samples (around 0.3 mg rhGH) weremixed with 0.5 g KBr using a mortar and pestle. After be-ing transferred into a stainless steel die, samples were pressedinto a disc using a vacuum press. Infrared (IR) spectra were col-lected on a Bomem MB-series spectrometer (Montreal, Provinceof Quebec, Canada). The spectra of the dried samples and ofnative aqueous rhGH were obtained as described previously.8

Data were processed to obtain second-derivative IR spectra ac-cording to a previous publication.8 Finally, for the major nega-tive band associated with the "-helix content of rhGH, the peakwidth at half height (w1/2) was computed by subtracting the lowwavenumber from the high one at the half peak height.8

Circular Dichroism Spectroscopic Measurement of rhGHSecondary Structure After Reconstitution

Circular dichroism (CD) spectra of rhGH in aqueous solutionbefore lyophilization and after reconstitution of lyophilizedsamples were obtained with a Chirascan-Plus (Applied Photo-physics, Leatherhead, Surrey, UK) CD spectrometer. For eachsample, the CD spectrum was plotted from 200 to 260 nm byaveraging spectra from triplicate measurements.8

Surface Area Measurement

A Quantachrome Autosorb-1 (Boynton Beach, Florida) was em-ployed to measure the SSAs of lyophilized formulations, usingfive-point krypton adsorption isotherms. For each formulation,samples from five vials of placebo formulation (no protein) wereplaced into the cell. The cell with samples was held under vac-uum for at least 5 h at room temperature to remove the mois-ture prior to initiation of the surface area measurement. Tripli-cate samples were measured to determine the surface area foreach sample.

Electron Spectroscopy for Chemical Analysis

To prevent from moisture uptake by dried formulations, allsample handling and preparation were performed in a glovebag purged with dry air (RH < 5%). Samples of lyophilizedpowders were deposited using double-sided adhesive tape ontoa sample holder (45◦) that was covered with a copper tape.The sample holder was transferred to the analysis chamberfor electron spectroscopy for chemical analysis (ESCA) mea-surements. ESCA was performed with a scanning auger multiprobe PHI spectrometer, model 25-120 (Physical ElectronicsInc., Eden Prairie, MN, US), equipped with monochromatic AlK" source (pass energy 100 eV). The C (1s) photoelectron line at284.6 eV was used as an internal standard for the correction ofthe charging effect in all samples. The vacuum was maintainedat approximately 10−8 Torr or lower. Spectra were collected byAugerScan (version 3.22) and analyzed using CasaXPS soft-ware (version 2.3.12) (Casa Software Ltd, Cheshire, UK). Atleast triplicate samples were used to determine elemental com-position for each formulation.

Calculation of the Mass of rhGH on the Surface of LyophilizedFormulations

Electron spectroscopy for chemical analysis was used to mea-sure elemental compositions of the surface layer of lyophilizedpowders. ESCA is sensitive to elemental composition in approx-imately the outermost 100 A of the lyophilized powders,15 andthus was used to measure the mass fraction of rhGH on the sur-face. The mass of final solid (mt) in each vial after lyophilizationis about 52 mg, which is essentially the sum of all the compo-nents: rhGH (1 mg), sugar (50 mg), phosphate buffer (∼0.5 mg),and residual water (less than 1% of total mass). rhGH containscarbon (C), oxygen (O), nitrogen (N), hydrogen (H), and sulfur(S), and based on its primary sequence, its atomic compositionis C995H1541N263O301S8. According to the manufacturer, ESCAis not capable of detecting H or any element whose surfaceconcentration is less than 0.1%. In the lyophilized solids thatcontain protein, sugar, and buffer, the overall S content is di-luted down to less than 0.1%; thus, all ESCA results reflectonly the surface elemental composition of N, C, and O. More-over, because neither sugar nor buffer contain N, the N peakis indicative of rhGH molecules in the outmost 100 A, that is,protein on the surface of the dried formulation. The calculatedtheoretical overall N content of rhGH on a sulfur- and H-freebasis is 18.0%. To calculate the mass of protein molecules onthe surface, we assume that the surface layer thickness probedby ESCA (l) is equal to 100 A.15 Furthermore, following theanalysis presented in earlier studies,9,15 we assumed that thedensity of the solid fraction within the cake (ρ) is constantacross the cake, with a value of roughly 1.1 g cm−3.9 Underthese assumptions, the mass fraction of rhGH in the surface

DOI 10.1002/jps.23926 Xu et al., JOURNAL OF PHARMACEUTICAL SCIENCES


layer is:

msurfaceprotein/mt = SSA ∗ lD ∗ N%/18.0% (1)

where mt (52 mg) is total lyophilizate mass per vial, SSA is eachformulation SSA per gram cake (m2/g), l (100 A) is the surfacethickness, and ρ (1.1 g/cm3) is the density of the solid withinthe cake. N% is the surface N percentage measured by ESCA.

rhGH Storage Stability Study

Lyophilized samples were incubated for 16 weeks at 323 K. Ateach time point (immediately after lyophilization, and after 1, 4,9, and 16 weeks), after rehydration of samples with water rhGHaggregation, deamidation, and oxidation levels were measuredby size-exclusion chromatography (SEC), ion-exchange chro-matography, and reverse-phase chromatography, respectively.8

At least triplicate measurements were carried out to determinethe quantities of remaining of monomer/native protein at eachtime point.

Error Analysis

Throughout the manuscript, when error bars are presented,they represent the experimental mean ± standard deviation,based on n ≥ 3.

RESULTS

Design of the Predrying Annealing Step

A requirement for predrying annealing is that the sample tem-perature should be maintained below the freezing point butabove the Tg

′ (or the eutectic melting temperature if there areany crystalline components). Tg

′ values determined from DSCexperiments were −15.5 ± 0.3◦C, −37.2 ± 0.2◦C, and −38.9 ±0.2◦C for 5% HES, 5% trehalose, and 5% sucrose solutions, re-spectively. In addition, DSC results showed that ice melt on-set temperature of these formulations was −2.5 ± 0.5◦C. Onthe basis of these results and previous successful annealingprotocols,11 a shelf temperature of −5◦C was selected for theannealing temperature for all three formulations.

Cake Structures, Water Content, and Glass TransitionTemperatures of Lyophilized Formulations

Formulations prepared by all five lyophilization cycles resultedin visually elegant cake structures.8 Water contents for alllyophilized samples prepared by all the different cycles wereless than 1% (w/w). The DSC measurements for all three for-mulations prepared by five different cycles showed single tran-sitions at their respective Tgs. Tgs for the three formulationsdid not show large variations as a function of the lyophiliza-tion cycle parameters. Tgs were 203 ± 3◦C, 98 ± 2◦C, and 64 ±2◦C for HES, trehalose and sucrose formulations, respectively(Table 1).28,29

Protein Structure in Lyophilized Formulations

Immediately after lyophilization, samples were analyzed withIR spectroscopy to obtain second-derivative spectra. The dom-inant negative band at 1654 cm−1 corresponds to the "-helical structure of rhGH.8 Figure 1 shows spectra for rhGHlyophilized in the presence of 5% HES (Fig. 1a), 5% trehalose(Fig. 1b), and 5% sucrose (Fig. 1c), for each of the different

Table 1. Tg (Onset) Temperatures Measured by DSC, PreviouslyReported ln(J$, h) Measured by Thermal Activity Monitor28 andPreviously Reported <u2>−1 Measured by Neutron Scattering29 forLyophilized Formulations (Without Protein) Prepared Using theStandard Freeze-Drying Cycle

Tg (Onset)(◦C) ± SD

ln(J$, h)(at 323 K)

<u2>−1 (A−2)± SD (at 323 K)

5% HES 203 ± 3 3.4 2.25 ± 0.045% Trehalose 98 ± 2 2.0 9.10 ± 0.505% Sucrose 64 ± 2 0.2 6.72 ± 0.17

lyophilization cycles tested. In general, based on the half-heightwidth and the depth of the negative of the "-helix band near1654 cm−1 measured by IR, rhGH in sucrose and trehalose for-mulations exhibited greater native-like secondary structuralcontent than rhGH lyophilized in HES formulations; sucroseformulations dried by all cycles led to the most native-likerhGH secondary structure. Most interesting, lyophilization cy-cles utilizing either predrying annealing or postdrying an-nealing yielded lyophilized formulations with more native-likerhGH secondary structure than those produced using the stan-dard lyophilization cycle. In contrast, fast freezing methods (N2

immersion and N2-droplet-freezing lyophilization) resulted inless native-like structures, as shown in Figure 1.

Protein Structure in Lyophilized and Reconstituted Formulations

After lyophilization, samples were immediately reconstitutedwith water, and the secondary structure of rhGH was measuredby far-UV (200–260 nm) CD spectroscopy (Fig. 2). The spectrafor the protein from all the formulations prepared by each of thedifferent lyophilization methods were indistinguishable fromthat of the aqueous native control protein. The CD data sug-gested that the secondary structural perturbations observed byIR in lyophilized samples were largely reversible upon recon-stitution.

SSA, Surface N Percentage and Amount of Protein on the Surface

Specific surface areass of the lyophilized formulations weremeasured using BET krypton adsorption (Table 2). The N2 im-mersion and N2-droplet-freezing lyophilized samples had muchlarger SSAs than the respective samples lyophilized using thestandard cycle. Postdrying annealing had minimal impact onthe SSAs of the lyophilized formulations we studied. On thecontrary, predrying annealing reduced SSAs up to twofold com-pared with standard lyophilized samples. Furthermore, regard-less of which lyophilization method was used, higher SSAs werefound for the HES formulation compared with those for the dis-accharide formulations.

Electron spectroscopy for chemical analysis measurementswere first performed on a lyophilized control sample contain-ing a 2:1 weight ratio of rhGH to phosphate buffer salts, with-out additional excipients. The control measurements yielded12.6 wt % N (on a sulfur- and H-free basis), which comparedwell with the expected theoretical value of 12 wt % N. ESCAmeasurements of the N% on the surface of the lyophilized cakescontaining glass-forming excipients are reported in Table 2.If the rhGH were homogeneously distributed throughout thedried solids, an N% value of 0.35% would be expected. How-ever, in all samples tested, the N% values measured in the out-ermost surface layer probed by ESCA were significantly higher,



Figure 1. Second-derivative IR spectroscopic analysis of freeze-driedformulations analyzed immediately after lyophilization. (a) 5% HESformulation, (b) 5% trehalose formulation, (c) 5% sucrose formulation.Formulations prepared by predrying annealing lyophilization (red),postdrying annealing (orange), standard lyophilization (purple), N2 im-mersion (green), N2 droplet freezing (blue), and aqueous native control(black).

indicating the presence of surface excesses of rhGH. Comparedwith results for the standard lyophilization cycle, both the twofast freezing methods (N2 immersion and N2 droplet freezing)showed slightly higher N% in the outmost 100 A of the driedformulations. On the contrary, postdrying annealing sampleshad low N%, and predrying annealing samples had N% valuesequivalent to those for standard lyophilized samples. For allprocessing methods, samples formulated in HES had higherlevels of surface N% than sucrose or trehalose formulations.

Figure 2. Protein secondary structures measured by CD spec-troscopy. Samples prepared by different lyophilization cycles were re-constituted immediately after lyophilization, and there were no signif-icant variations in the CD spectra between different samples.

Finally, the total amount of rhGH in the outmost 100 A layerwas calculated using Eq. (1) (Table 2). Regardless of the type ofexcipients used, samples prepared by two rapid freezing, liquidN2 treatment methods showed the highest amounts of proteinon their surfaces, as a result of both larger SSAs and highersurface N percentages. Likewise, regardless of the processingmethod that was used, the amounts of rhGH on the surface oflyophilized HES formulations were higher than those for for-mulations prepared from sucrose or trehalose solutions.

Lyophilization Impact on rhGH Monomer Levels

Prior to lyophilization, SEC analysis of rhGH showed that theprotein was more than 99.9% monomeric. Upon reconstitu-tion of samples immediately after lyophilization, about 2%–3% monomer loss (compared with control samples) was ob-served in all formulations prepared by the two faster freezingmethods (N2 immersion and N2-droplet-freezing lyophilization

Table 2. Specific Surface Areas of Lyophilized Formulations, AtomPercentage N in the Outmost 100 A Layer of LyophilizedFormulations and the Mass Fraction of Protein in that Layer

SSA (m2/g) N (%)Surface rhGHFraction (%)

Standardlyophilization

5% HES 2.3 ± 0.2 1.4 ± 0.2 10.2

5% Trehalose 1.6 ± 0.1 0.6 ± 0.1 3.15% Sucrose 1.2 ± 0.1 0.7 ± 0.1 2.7

Predryingannealing

5% HES 1.8 ± 0.1 1.4 ± 0.1 8.0


Postdryingannealing

5% HES 2.3 ± 0.2 1.2 ± 0.2 8.8

5% Trehalose 1.3 ± 0.1 0.4 ± 0 1.75% Sucrose 1.2 ± 0.0 0.5 ± 0.1 1.9

N2 immersion 5% HES 6.0 ± 0.5 1.5 ± 0.2 28.65% Trehalose 2.6 ± 0.3 0.9 ± 0.1 7.55% Sucrose 2.7 ± 0.3 0.8 ± 0.1 7.0

N2 dropletfreezing

5% HES 6.2 ± 0.5 1.6 ± 0.2 31.5




Figure 3. Percentage of rhGH monomer detected by SEC analysisafter lyophilization with various cycles and immediate reconstitution.( ) Standard lyophilization; ( ) predrying annealing; ( ) postdry-ing annealing; ( ) N2 immersion; ( ) N2 droplet freezing. Relativelyhigher monomer loss was seen in the samples prepared by N2 immer-sion and N2-droplet-freezing lyophilization methods.

cycles) (Fig. 3). On the contrary, no significant differences wereobserved in levels of deamidation or oxidation before and af-ter lyophilization process, regardless of which formulation orlyophilization cycle was used (data not shown).

Storage Stability of rhGH Lyophilized Samples

The percentages of remaining monomeric rhGH, unoxidizedrhGH, and nondeamidated rhGH were determined at each timepoint throughout the 16-week incubation study (Fig. 4). Therate of loss of native protein because of the formation of degra-dation products decreased with increasing time. Some formula-tions showed plateaus by the end of 16-week incubation study.It is clear that among all the three formulations, the highestlevels of damage were seen in the samples prepared by fastfreezing methods. In contrast, samples treated with annealing(predrying or postdying annealing) showed least damage afterthe 16-week incubation period. In addition, both deamidationand oxidation were faster than aggregation, and disaccharideformulations (sucrose and trehalose) exhibited slower degrada-tion kinetics for rhGH than HES formulations did.

DISCUSSION

We hypothesized that the dominant factor that determinesthe rate of protein degradation observed during storage oflyophilized formulations is the amount of protein found at thesolid–air interface after lyophilization. By employing differentglass-forming stabilizers and lyophilization methods, sampleswith a wide range of masses of protein on solid–air interfacewere generated as a result of modulating both the SSA of theglass and concentration of the protein in the surface layer.

One major factor contributing to differences in the SSAsobserved with the various lyophilization processes is Ostwaldripening of ice crystals, which only occurs to a significant ex-tent between the time when freezing is initiated and the timewhen the samples cool down to the glass transition tempera-

ture for the maximally freeze-concentrated solution (Tg′). Sam-

ples prepared by the predrying annealing method (wherein thetemperature was maintained between the freezing tempera-ture and Tg

′) had the smallest SSA values of all the methodstested because the time available for Ostwald ripening of icecrystals was longest in that method. In contrast, the fast freez-ing methods (liquid N2 immersion or droplet freezing) causethe solution temperature to rapidly reach Tg

′, thus allowinglittle time for Ostwald ripening and yielding correspondinglyhigh SSA values. Finally, we observed that the SSA values forthe formulation processed with the postdrying annealing cyclewere equivalent to those for the standard cycle. This can be ex-plained because the freezing portion of the lyophilization cyclewas identical for the standard and postdrying annealing meth-ods. Hence, the times available for Ostwald ripening for the twomethods were equal, resulting in equal ice crystal size distri-butions and equivalent SSA values. Also, as long as postdryingannealing is executed well below the system Tg, no change inSSA would be expected during this process.

The N% of the various lyophilized formulations determinedby ESCA is indicative of protein fraction on the surface, be-cause no other formulation component contains N. For a givenformulation, N% was similar for samples prepared by predryingannealing, postdrying annealing, and standard lyophilizationmethods, consistent with previous studies on methionyl rhGH.9

Samples prepared by the two liquid N2-treated lyophilizationprotocols resulted in higher N% (Table 2). In addition, the HESformulations had the highest protein fraction on the surface re-gardless of which lyophilization method was used, again consis-tent with the earlier study on methionyl rhGH, which demon-strated that surface concentrations of protein were highest informulations containing polymeric excipients.9 A possible ex-planation for the higher masses of protein found on the surfaceof lyophilized HES formulations is that, during the freezingstep of the lyophilization cycles, the greater viscosity of HESformulations hindered the diffusion of the protein away fromgrowing ice crystal surfaces. The explanation for the high N%found on the surface of formulations prepared with liquid N2

immersion or droplet freezing may be similar. The rapid iceformation in the process and short time between initiation offreezing and sample glassification at Tg

′ limited the time avail-able for protein to diffuse away from growing ice surfaces.

Webb et al.19 reported that lyophilized formulations of recom-binant human interferon-( with higher SSAs also had higherrates of protein aggregation. It was also noted, based on ESCA9

measurements, that the surfaces of the lyophilized solids wereenriched in protein. This enrichment was likely because of acombination protein adsorption at ice–water interfaces, andlimited ability of large molecules such as proteins to diffuseaway from the freezing front causing them to be trapped at thesurface.10 Consistent with previous reports,15,30,31 we observedsubstantial enrichment of protein on the solid–air surface oflyophilized powders (Table 2). Proteins found on the solid–airinterfaces of glassy lyophilized solids certainly experience anenvironment that is dramatically different from that inside theglass.32 Effective glass transition temperatures are expected tobe lower in the interfacial region than in the bulk,33,34 allowinggreater mobility for any protein molecules found at the surfaceof glassy lyophilized powders.

Recombinant human growth hormone structures in driedsolids prepared using both predrying annealing and postdry-ing annealing were more native-like than in those prepared



Figure 4. Storage stability as a function of incubation time at 323 K for rhGH in formulations lyophilized using various cycle parameters. (a)Fraction monomeric rhGH remaining after storage in glassy HES; (b) fraction rhGH not deamidated after storage in glassy HES; (c) fractionrhGH that is not oxidized after storage in glassy HES; (d) fraction monomeric rhGH remaining after storage in glassy trehalose; (e) fractionrhGH not deamidated after storage in glassy trehalose; (f) fraction rhGH that is not oxidized after storage in glassy trehalose; (g) fractionmonomeric rhGH remaining after storage in glassy sucrose; (h) fraction rhGH not deamidated after storage in glassy sucrose; (i) fraction rhGHthat is not oxidized after storage in glassy sucrose. For all panels, lyophilization cycle conditions are represented by �: standard lyophilization,•: predrying annealing, �: postdrying annealing, �: N2 immersion, �: N2 droplet freezing. Panels a–c represent HES formulation, panels d–frepresent trehalose formulation, and panels g–i represent sucrose formulation. Connected lines are predicted values from two parameter (ks andkb) first-order kinetics model for each individual formulation, using surface protein quantities determined from ESCA and SSA measurements.

using the standard lyophilization cycle, whereas rhGH struc-tures were most perturbed in samples that had been rapidlyfrozen using liquid N2. Webb et al.19 proposed that annealingcould serve to alleviate residual stress and reduce the excessfree volume of the glass, thus improving protein structures.However, another explanation could be that protein moleculeson the solid–air surface are more prone to structural perturba-tion. Hence, those samples with smaller quantities of protein atthe solid–air surface (i.e., predrying annealed samples) shouldshow more retention of native structure, whereas samples withlarger surface protein quantities (e.g., those prepared with liq-uid N2 freezing) should show more perturbed structures. This isindeed the case, as can be seen in Figure 5, where formulationsexhibiting a larger fraction of rhGH at the interface had lessnative-like structure as measured by IR, that is, larger valuesof �w1/2.8

In addition, previously, we showed that decreased activa-tion free energies (�G†) for aggregation, deamidation, and ox-idation of rhGH correlated with increasing degrees of protein

Figure 5. Change of IR "-helical peak half width measured inlyophilized solids compared with that of native rhGH in aqueous so-lution (�w1/2), plotted against the fraction of the total protein found onsurface. For some data points, error bars are smaller than the symbols.



Figure 6. Correlation of the percent of rhGH found as aggregatesmeasured after 16 weeks of incubation at 323 K with the change com-pared with native rhGH in aqueous solution of the IR "-helical peakhalf width in the lyophilized solid formulations (�w1/2). For some datapoints, error bars are smaller than the symbols.

structural perturbation in lyophilized formulations.8 In turn,rhGH degradation (by aggregation, oxidation, and deamida-tion) was faster in samples wherein the rhGH structure wasmore perturbed.8 A similar result was seen in this study. Asshown in Figure 6, the fraction of aggregated protein after16 weeks of incubation at 323 K increases with the degreeof rhGH structural perturbation, as reflected in the change af-ter lyophilization of the IR "-helical peak width at half height,�w1/2.

Kinetic Model for rhGH Degradation in Lyophilized Samples

Because populations of protein molecules found at the solid–air interface of lyophilized samples are likely to experience asignificantly different environment from those molecules foundin the bulk, we analyzed the aggregation, deamidation, andoxidation kinetics for rhGH by assuming simple first-order de-pendence of degradation kinetics, both in the surface layer andin the bulk. Thus, for each of the degradation reactions, wewrite:

Pst(t) = Psoe−ks,it (2)

Pbt(t) = Pboe−kb,it (3)

Ptot(t) = Pst(t) + Pbt(t) (4)

where Pst and Pbt are the amount of native protein on the sur-face and in the bulk, respectively, at a certain time t; Pso andPbo are the initial amount of native protein on the solid surfaceand in the bulk, respectively; and ks,i and kb,i are the apparentfirst-order degradation rate constant for protein on the surfaceand in bulk, for each of the i reactions (aggregation, deamida-tion, and oxidation). The sum of Pst and Pbt is equal to Ptot, theamount of remaining native protein at any given time point,which can be measured by size-exclusion, ion-exchange, andreverse-phase chromatographic analysis of the samples afterreconstitution. For each degradation pathway and each formu-lation, using initial amounts of protein on the surface and in

the bulk determined from the ESCA and SSA measurements forPso and Pbo, the two parameters ks,i and kb,i were fit to the datafrom the 16-week incubation study (Table 3), using the evolu-tionary solving method in the Microsoft Excel R© solver package,with constraints of convergence as 10−8, mutation rate as 0.9,and maximum time as 120 s. Convergence to the reported op-timal values was obtained from multiple initial guess values.Predicted kinetics based upon ks,i and kb,i determined for eachformulation is plotted in Figure 4. In general, predicted valuesfrom individual two parameter (ks,i and kb,i) first-order kineticsmodel fit well to the experimental data.

As expected, values of the surface-layer rate constants ks,i

were much higher than the bulk glass rate constants kb,i. Foraggregation, the rate constant in the bulk ks,agg was negligible,as expected because of the high viscosity in the glassy state thatlimits diffusive transport of large molecules such as rhGH andalso restricts the relatively large-scale motions required for theprotein unfolding that is generally associated with aggregation.Oxidation and deamidation reactions, both of which requiresmaller degrees of molecular motion than does aggregation,4

showed rate constants in the bulk glass that were roughly twoorders of magnitude smaller than those observed for rhGH inthe surface layer.

A striking result of the analysis is that the fitted rate con-stants do not depend on the composition of the formulationsor on the lyophilization cycle that was used to generate thesamples. Although parameters that reflect molecular motionswithin the glassy formulations such as the relaxation timeJ$ determined from thermal activity monitor measurements28

and the inverse mean-square displacement of hydrogen atoms<u2>−1 (see Table 1) are different for the three formulationswe examine here, the rate constants for the three reactionsare insensitive to these glassy-state relaxation properties (wenote that the relaxation properties were measured in protein-free samples and may not reflect motions within the proteinmolecules themselves.). In fact, the respective surface and bulkrate constants for the various formulations and lyophilizationcycle parameters are so similar that the data for each typeof rhGH degradation during incubation at 323 K could be fitusing Eqs. (2–4) by two global formulation and lyophilizationcycle independent, apparent first-order rate constants. Table 4lists the three sets of global rate constants ksg,i and kbg,i, wherethe subscript i refers to the degradation pathway (aggregation,deamidation, or oxidation).

Figure 7 shows the experimentally determined fractions ofthe rhGH in the various lyophilized formulations that weredamaged by aggregation, oxidation, and deamidation after 16weeks of storage at 323 K, plotted against values predictedfrom Eqs. (2–4), measured values of the quantity of rhGH inthe surface layer and the fitted rate constants presented inTable 4. The simple model correlates the data very well, witha regression line for the plot of actual versus predicted extentof degradation yielding a slope of 1.07 ± 0.02 and a correlationcoefficient R2 = 0.98.

Proteins lyophilized in glassy formulations often show degra-dation as a function of storage time.4,8 This observation is per-plexing, especially for protein aggregation, because the highviscosities (>1013 poise35) found in the glassy state shouldpreclude the relatively large-scale diffusive motions requiredfor proteins to aggregate. One possible explanation to this co-nundrum is that during storage, protein molecules in glassysolids might accumulate small conformational changes that



Table 3. Apparent First-Order Rate Constants for rhGH Degradation During Incubation at 323 K in the Surface Layer (ks,i) and in theGlassy Bulk Solid Portion (kb,i) for Each Formulation and Lyophilization Cycle

Apparent First-Order Rate Constants (% Per Week)

Route of Degradation Formulations Standard Preannealing Postannealing N2 Immersion N2 Droplet Freezing

Aggregation ks,agg kb,agg ks,agg kb,agg ks,agg kb,agg ks,agg kb,agg ks,agg kb,agg

5% HES 12.1 0 10.7 0 9.5 0 12.7 0 13.3 05% Trehalose 5.7 0 6.5 0 6.6 0 4.5 0 4.6 05% Sucrose 6.6 0 4.4 0 4.5 0 3.8 0 4.2 0

Standard Preannealing Postannealing N2 Immersion N2 Droplet Freezing

Deamidation ks,d kb,d ks,d kb.d ks.d kb,d ks,d kb,d ks,d kb,d

5% HES 32.0 0.7 30.9 0.5 29.8 0.4 33.2 0.5 36.4 0.65% Trehalose 32.2 0.6 28.4 0.3 26.4 0.4 29.4 0.3 19.0 0.55% Sucrose 30.6 0.7 23.4 0.4 32.6 0.5 31.2 0.5 28.6 0.5

Standard Preannealing Postannealing N2 Immersion N2 Droplet Freezing

Oxidation ks,ox kb,ox ks,ox kb,ox ks,ox kb,ox ks,ox kb,ox ks,ox kb,ox

5% HES 92.9 0.8 99.3 0.7 97.6 0.5 92.4 0.8 93.6 0.95% Trehalose 88.7 0.6 85.8 0.6 77.8 0.5 95.1 0.7 86.8 0.85% Sucrose 81.5 0.8 81.5 0.5 83.3 0.6 83.7 0.7 69.2 0.7

The subscript i refers to the type of degradation: aggregation (i = agg), deamidation (i = d) or oxidation (i = ox). Mean absolute percent deviation for these fitsare all below 8%.

Table 4. Global First-Order Degradation Rate Constants at 323 Kfor Each Degradation Pathway (Aggregation, Deamidation, andOxidation) for rhGH on the Surface and Within the Bulk Solid of Allthe Formulations

Degradation Route ksg,i (% Per Week) kbg,i (% Per Week)

i = aggregation 11.6 0 (<10−3)i = deamidation 35.7 0.5i = oxidation 90.8 0.7

prime them to become increasingly aggregation-competentupon reconstitution.4,5 However, no evidence showing the ac-cumulation of this type of aggregate-competent species duringstorage has been presented,36 perhaps in part because of therelatively low sensitivity of available optical spectroscopies toexamine protein structure in dry solids.37 Aggregation of pro-teins after lyophilization and reconstitution is often correlatedwith the extent of loss of native protein structure that occursduring the lyophilization process, with those formulations andconditions that yield the greatest loss of native structure re-sulting in the most aggregation upon reconstitution.8 Althoughnumerous studies38–47 have demonstrated the role of unfoldedor partially unfolded protein molecules in fostering aggrega-tion, the acute loss of structure during lyophilization does notcompletely explain why aggregation levels increases duringstorage.

In current work, by separating the degradation kinetics be-tween surface and bulk of the solid, we find an alternativeexplanation for protein aggregation during storage. The ma-jority of the degradation occurs in the population of proteinmolecules found on the solid surface. Adsorption of proteinsto interfaces frequently results in conformational perturba-tions, gelation, and aggregation, for example.48–54 Furthermore,in thin films of poly(methyl methacrylate)55 and polystyrene56

Figure 7. Comparison of the predicted fraction protein damaged after16 weeks of incubation at 323 K with experimental values measuredby liquid chromatography methods (error bars on experimental valuesare presented in Fig. 4 and omitted here for clarity). The predictedfraction was computed using pairs of global degradation rate constantsksg,i and kbg,i. The degradation routes are denoted •: i, aggregation; �:i, deamidation; �: i, oxidation). The linear regression line has a slopeof 1.07 ± 0.02, R2 = 0.98.

formed by drying of spin-coated layers, residual stress remain-ing in the surface layer causes slow relaxation motions andconformational rearrangement of polymer chains. It is plausi-ble that the correlation that we observe between the fractionsof the rhGH molecules found in the surface layer of lyophilizedpowders and the loss of native protein secondary structure (seeFig. 5) is because of a similar effect, wherein protein moleculesin the glass–ice surface layer experience residual stress upondrying, causing the observed unfolding. Furthermore, the glasstransition temperature near the surface of the glasses varies



from that of the bulk. For example, glass transition tempera-tures found in a surface layer approximately 1000 A thick inpoly(methyl methacrylate) films differ from those of the bulk.57

Yu and coworkers32,58 reported that diffusion on the surface ofglasses was at least 106 times faster than bulk diffusion, andglass surface diffusion can cause surface evolution at nm to :mscale, which is a length scale that would be expected to be longenough for (conformationally perturbed) protein molecules tocollide with each other in the surface glass layer, causing ag-gregation.

CONCLUSIONS

Predrying annealing and postdrying annealing both resultmore native-like rhGH structure after lyophilization. Fast-freezing lyophilization cycle is detrimental for rhGH not onlyduring lyophilization process, but also during storage. Theamount of protein on the solid–air interface should be a keyfactor to consider for formulation and lyophilization design, asprotein on the surface degrades much faster than that in thebulk. Because of the substantial mobility differences betweenproteins on the glass surface and in the bulk, calculating bothsurface and bulk degradation kinetics is recommended to ra-tionally design protein formulation and lyophilization process.

ACKNOWLEDGMENT

We acknowledge the funding from NIH/NIBIB under grantnumber R01 EB006398-01A1.

REFERENCES

1. Carpenter JF, Pikal MJ, Chang BS, Randolph TW. 1997. Rationaldesign of stable lyophilized protein formulations: Some practical advice.Pharm Res 14(8):969–975.2. Manning MC, Patel K, Borchardt RT. 1989. Stability of protein phar-maceuticals. Pharm Res 6(11):903–918.3. Wang W. 2000. Lyophilization and development of solid protein phar-maceuticals. Int J Pharm 203(1–2):1–60.4. Chang LQ, Pikal MJ. 2009. Mechanisms of protein stabilization inthe solid state. J Pharm Sci 98(9):2886–2908.5. Cicerone MT, Douglas JF. 2012. beta-Relaxation governs proteinstability in sugar–glass matrices. Soft Matter 8(10):2983–2991.6. Carpenter JF, Prestrelski SJ, Dong AC. 1998. Application of infraredspectroscopy to development of stable lyophilized protein formulations.Eur J Pharm Biopharm 45(3):231–238.7. Randolph TW. 1997. Phase separation of excipients duringlyophilization: Effects on protein stability. J Pharm Sci 86(11):1198–1203.8. Xu Y, Carpenter JF, Cicerone MT, Randolph TW. 2013. Contribu-tions of local mobility and degree of retention of native secondary struc-ture to the stability of recombinant human growth hormone (rhGH) inglassy lyophilized formulations. Soft Matter 9(32):7855–7865.9. Abdul-Fattah AM, Lechuga-Ballesteros D, Kalonia DS, Pikal MJ.2008. The impact of drying method and formulation on the physicalproperties and stability of methionyl human growth hormone in theamorphous solid state. J Pharm Sci 97(1):163–184.10. Kasper JC, Friess W. 2011. The freezing step in lyophilization:Physico-chemical fundamentals, freezing methods and consequenceson process performance and quality attributes of biopharmaceuticals.Eur J Pharm Biopharm 78(2):248–263.11. Searles JA, Carpenter JF, Randolph TW. 2001. Annealing to op-timize the primary drying rate, reduce freezing-induced drying rate

heterogeneity, and determine Tg′ in pharmaceutical lyophilization. J

Pharm Sci 90(7):872–887.12. Rathore N, Rajan RS. 2008. Current perspectives on stability ofprotein drug products during formulation, fill and finish operations.Biotechnol Prog 24(3):504–514.13. Heller MC, Carpenter JF, Randolph TW. 1999. Protein formulationand lyophilization cycle design: Prevention of damage due to freeze-concentration induced phase separation. Biotechnol Bioeng 63(2):166–174.14. Searles JA, Carpenter JF, Randolph TW. 2001. The ice nucle-ation temperature determines the primary drying rate of lyophiliza-tion for samples frozen on a temperature-controlled shelf. J Pharm Sci90(7):860–871.15. Webb SD, Golledge SL, Cleland JL, Carpenter JF, RandolphTW. 2002. Surface adsorption of recombinant human interferon-gamma in lyophilized and spray-lyophilized formulations. J Pharm Sci91(6):1474–1487.16. Patel S, Bhugra C, Pikal M. 2009. Reduced pressure ice fog tech-nique for controlled ice nucleation during freeze-drying. AAPS Pharm-SciTech 10(4):1406–1411.17. Cochran T, Nail SL. 2009. Ice nucleation temperature influencesrecovery of activity of a model protein after freeze drying. J Pharm Sci98(9):3495–3498.18. Rambhatla S, Ramot R, Bhugra C, Pikal MJ. 2004. Heat and masstransfer scale-up issues during freeze drying: II. Control and character-ization of the degree of supercooling. AAPS PharmSciTech 5(4):54–62.19. Webb SD, Cleland JL, Carpenter JF, Randolph TW. 2003. Effectsof annealing lyophilized and spray-lyophilized formulations of recom-binant human interferon-gamma. J Pharm Sci 92(4):715–729.20. Hsu CC, Nguyen HM, Yeung DA, Brooks DA, Koe GS, Bewley TA,Pearlman R. 1995. Surface denaturation at solid–void interface—A pos-sible pathway by which opalescent particulates form during the storageof lyophilized tissue-type plasminogen-activator at high-temperatures.Pharm Res 12(1):69–77.21. Patapoff TW, Overcashier DE. 2002. The importance of freezing onlyophilization cycle development. BioPharm 15(3):16–72.22. Bhatnagar BS, Bogner RH, Pikal MJ. 2007. Protein stability duringfreezing: Separation of stresses and mechanisms of protein stabiliza-tion. Pharm Dev Technol 12(5):505–523.23. Chang BS, Kendrick BS, Carpenter JF. 1996. Surface-induced de-naturation of proteins during freezing and its inhibition by surfactants.J Pharm Sci 85(12):1325–1330.24. Maa YF, Nguyen PA, Sweeney T, Shire SJ, Hsu CC. 1999. Proteininhalation powders: Spray drying vs spray freeze drying. Pharm Res16(2):249–254.25. Wang B, Cicerone MT, Aso Y, Pikal MJ. 2010. The impact of thermaltreatment on the stability of freeze-dried amorphous pharmaceuticals:II. Aggregation in an IgG1 fusion protein. J Pharm Sci 99(2):683–700.26. Crisman RL, Randolph TW. 2010. Crystallization of recombi-nant human growth hormone at elevated pressures: Pressure ef-fects on PEG-induced volume exclusion interactions. Biotechnol Bioeng107(4):663–672.27. May JC, Grim E, Wheeler RM, West J. 1982. Determination of resid-ual moisture in freeze-dried viral vaccines—Fischer, Karl, gravimetricand thermogravimetric methodologies. J Biol Standard 10(3):249–259.28. Chieng N, Mizuno M, Pikal M. 2013. Characterization of dynam-ics in complex lyophilized formulations: I. Comparison of relaxationtimes measured by isothermal calorimetry with data estimated fromthe width of the glass transition temperature region. Eur J PharmBiopharm 85(2):189–196.29. Chieng N, Cicerone MT, Zhong Q, Liu M, Pikal MJ. 2013. Charac-terization of dynamics in complex lyophilized formulations: II. Analysisof density variations in terms of glass dynamics and comparisons withglobal mobility, fast dynamics, and positron annihilation lifetime spec-troscopy (PALS). Eur J Pharm Biopharm 85(2):197–206.30. Adler M, Unger M, Lee G. 2000. Surface composition of spray-dried particles of bovine serum albumin/trehalose/surfactant. PharmRes 17(7):863–870.



31. Millqvist-Fureby A, Malmsten M, Bergenstahl B. 1999. Spray-drying of trypsin—Surface characterisation and activity preservation.Int J Pharm 188(2):243–253.32. Zhu L, Brian CW, Swallen SF, Straus PT, Ediger MD, YuL. 2011. Surface self-diffusion of an organic glass. Phys Rev Lett106(25):256103.33. Zuo B, Liu YJ, Wang L, Zhu YM, Wang YF, Wang XP. 2013. Depthprofile of the segmental dynamics at a poly(methyl methacrylate) filmsurface. Soft Matter 9(39):9376–9384.34. Liu D, Orozco RO, Wang T. 2013. Deviations of the glass transitiontemperature in amorphous conjugated polymer thin films. Phys Rev E88(2):022601.35. Angell CA. 1995. Formation of glasses from liquids and biopolymers.Science 267(5206):1924–1935.36. Kaminski K, Adrjanowicz K, Zakowiecki D, Kaminska E, Wlodar-czyk P, Paluch M, Pilch J, Tarnacka M. 2012. Dielectric studies onmolecular dynamics of two important disaccharides: Sucrose and tre-halose. Mol Pharm 9(6):1559–1569.37. Jiang Y, Li C, Nguyen X, Muzammil S, Towers E, GabrielsonJ, Narhi L. 2011. Qualification of FTIR spectroscopic method forprotein secondary structural analysis. J Pharm Sci 100(11):4631–4641.38. Khurana R, Gillespie JR, Talapatra A, Minert LJ, Ionescu-ZanettiC, Millett I, Fink AL. 2001. Partially folded intermediates as criticalprecursors of light chain amyloid fibrils and amorphous aggregates.Biochemistry 40(12):3525–3535.39. Fink AL. 1998. Protein aggregation: Folding aggregates, inclusionbodies and amyloid. Fold Des 3(1):R9–R23.40. Fink AL, Calciano LJ, Goto Y, Kurotsu T, Palleros DR. 1994. Clas-sification of acid denaturation of proteins: Intermediates and unfoldedstates. Biochemistry 33(41):12504–12511.41. Carrotta R, Bauer R, Waninge R, Rischel C. 2001. Confor-mational characterization of oligomeric intermediates and aggre-gates in beta-lactoglobulin heat aggregation. Protein Sci 10(7):1312–1318.42. Gomez-Orellana I, Variano B, Miura-Fraboni J, Milstein S, PatonDR. 1998. Thermodynamic characterization of an intermediate state ofhuman growth hormone. Protein Sci 7(6):1352–1358.43. Speed MA, Morshead T, Wang DI, King J. 1997. Conformation ofP22 tailspike folding and aggregation intermediates probed by mono-clonal antibodies. Protein Sci 6(1):99–108.44. King J, Haase-Pettingell C, Robinson AS, Speed M, Mitraki A.1996. Thermolabile folding intermediates: Inclusion body precursorsand chaperonin substrates. FASEB J 10(1):57–66.

45. Kim D, Yu MH. 1996. Folding pathway of human alpha 1-antitrypsin: Characterization of an intermediate that is active butprone to aggregation. Biochem Biophys Res Commun 226(2):378–384.46. Grillo AO, Edwards KL, Kashi RS, Shipley KM, Hu L, BesmanMJ, Middaugh CR. 2001. Conformational origin of the aggregation ofrecombinant human factor VIII. Biochemistry 40(2):586–595.47. Chi EY, Krishnan S, Randolph TW, Carpenter JF. 2003. Physicalstability of proteins in aqueous solution: Mechanism and driving forcesin nonnative protein aggregation. Pharm Res 20(9):1325–1336.48. Bantchev GB, Schwartz DK. 2003. Surface shear rheology ofbeta-casein layers at the air/solution interface: Formation of a two-dimensional physical gel. Langmuir 19(7):2673–2682.49. Dickinson E. 1999. Adsorbed protein layers at fluid interfaces: In-teractions, structure and surface rheology. Colloids Surf B 15:161–176.50. Kragel J, Derkatch SR, Miller R. 2008. Interfacial shear rheologyof protein–surfactant layers. Adv Colloid Interface Sci 144(1–2):38–53.51. Vessely CR, Carpenter JF, Schwartz DK. 2005. Calcium-inducedchanges to the molecular conformation and aggregate structure of beta-casein at the air–water interface. Biomacromolecules 6(6):3334–3344.52. Britt KA, Schwartz DK, Wurth C, Mahler HC, Carpenter JF, Ran-dolph TW. 2012. Excipient effects on humanized monoclonal antibodyinteractions with silicone oil emulsions. J Pharm Sci 101(12):4419–4432.53. Bee JS, Chiu D, Sawicki S, Stevenson JL, Chatterjee K, Freund E,Carpenter JF, Randolph TW. 2009. Monoclonal antibody interactionswith micro- and nanoparticles: Adsorption, aggregation and acceleratedstress stability studies. J Pharm Sci 98(9):11.54. Bee JS, Randolph TW, Carpenter JF, Bishop SM, Dimitrova MN.2011. Effects of surfaces and leachables on the stability of biopharma-ceuticals. J Pharm Sci 100(10):4158–4170.55. Reiter G, Hamieh M, Damman P, Sclavons S, Gabriele S, Vilmin T,Raphael E. 2005. Residual stresses in thin polymer films cause ruptureand dominate early stages of dewetting. Nat Mater 4(10):754–758.56. Gabriele S, Damman P, Sclavons S, Desprez S, Coppee S, ReiterG, Hamieh M, Al Akhrass S, Vilmin T, Raphael E. 2006. Viscoelasticdewetting of constrained polymer thin films. J Polym Sci B 44(20):3022–3030.57. Inoue R, Nakamura M, Matsui K, Kanaya T, Nishida K, HinoM. 2013. Distribution of glass transition temperature in multilayeredpoly(methyl methacrylate) thin film supported on a Si substrate asstudied by neutron reflectivity. Phys Rev E 88(3):032601.58. Sun Y, Zhu L, Wu T, Cai T, Gunn EM, Yu L. 2012. Stability of amor-phous pharmaceutical solids: Crystal growth mechanisms and effect ofpolymer additives. AAPS J 14(3):380–388.


protein quantity at the air-solid interface publication

Documents