PHARMACEUTICAL TECHNOLOGY
Water Sorption, Glass Transition, and Protein-Stabilizing
Behavior of an Amorphous Sucrose Matrix Combined With
Various Materials
KOREYOSHI IMAMURA, TORU YOKOYAMA, ATSUSHI FUKUSHIMA, MITSUNORI KINUHATA, KAZUHIRO NAKANISHI
Department of Bioscience and Biotechnology, Faculty of Engineering, Okayama University, Okayama 700-8530, Japan
Received 3 December 2009; revised 19 February 2010; accepted 25 February 2010
Published online 13 May 2010 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.22162
ABSTRACT: The effects of various additives on the physical properties of an amorphous sugar
matrix were compared. Amorphous, sugar–additive mixtures were prepared by freeze-drying
and then rehumidified at given RHs. Sucrose and eighteen types of substances were used as the
sugar and the additive, respectively, and water sorption, glass-to-rubber transition, and protein
stabilization during freeze-drying for the various sucrose–additive mixtures were examined.
The additives were categorized into two groups according to their effects on Tg and water
sorption. Presence of polysaccharides, cyclodextrins, and polymers (large-sized additives)
resulted in a decrease in equilibrium water content from the ideal value calculated from
individual water contents for sucrose and additive, and in contrast, low MW substances
containing ionizable groups (small-ionized additives) resulted in an increase. The increase in
Tg by the addition of large-sized additives was significant at the additive contents >50 wt.%
whereas the Tg was markedly increased in the lower additive content by the addition of smallionized additives. The addition of small-ionized additives enhanced the decrease in Tg with
increasing water content. The protein stabilizing effect was decreased with increasing additive
content in the cases of the both groups of the additives. ß 2010 Wiley-Liss, Inc. and the American
Pharmacists Association J Pharm Sci 99:4669–4677, 2010
Keywords:
stabilization
amorphous; sugar; additives; water sorption; glass transition; protein
INTRODUCTION
An amorphous matrix, comprised of sugar, is
frequently used as a stabilizing agent for proteins
that are unstable against dehydration and storage in
the drug manufacturing industries.1–3 In general,
low-molecular-weight sugars such as disaccharides,
are thought to be superior in stabilizing proteins, as
opposed to larger oligosaccharides.4–6 However,
amorphous matrices of low-molecular-weight sugars
are likely to show a glass-to-rubber transition by a
increase in temperature and a slight water sorption7
and subsequently lose their stabilizing effects.8,9 As a
result, improving the physical stability of amorphous
Additional Supporting Information may be found in the online
version of this article.
Correspondence to: Koreyoshi Imamura (Telephone: 086-251-8201;
Fax: 086-251-8201; E-mail: [email protected])
Journal of Pharmaceutical Sciences, Vol. 99, 4669–4677 (2010)
ß 2010 Wiley-Liss, Inc. and the American Pharmacists Association
matrices that are made up of low-molecular-weight
sugars is an important issue.5,6
One of the methods used to improve the physical
stability of amorphous mono- and disaccharides is to
add a substance that can cause an increase in the
glass transition temperature (Tg) to the matrix of lowmolecular-weight sugars.5,6,10 Polysaccharides and
polymers have been reported to increase the Tg values
of amorphous oligosaccharides.11–14 On the other
hand, it has recently been reported that the addition
of phosphate salts to an amorphous sugar matrix can
also increase the Tg value.15–18 This suggests that
highly polarized substances and substances that
contain polar groups could also be used to improve
the physical stability of an amorphous sugar matrix.
The physical stability of an amorphous sugar
matrix is generally known to be quite sensitive to
the amount of water sorbed to the matrix.19 The water
sorption behavior of an amorphous sugar matrix
results in significant change in the presence of other
components,13,20,21 which would naturally be expect-
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010
4669
4670
IMAMURA ET AL.
ed to have a potent effect on the dependence of Tg on
the amount of sorbed water. Such changes in the
physical properties of an amorphous sugar matrix
would also be expected to affect the protein-stabilizing effect of the matrix. Accordingly, water sorption,
glass-to-rubber transition, and the protein stabilizing
effect of a sugar–additive binary system, should be
comprehensively investigated to understand the
impact of additives on an amorphous sugar matrix.
However, to our knowledge, such an investigation has
not been reported.
In this study, various additives were compared for
their effects on water sorption behavior, Tg, and the
protein-stabilizing effect of an amorphous sugar
matrix. Sucrose was used as a model sugar, and 18
types of substances, including polysaccharides, polymers, salts, and glucosides containing ionizable
groups, were used as additives. Various compositions
of amorphous sucrose–additive mixtures were prepared by freeze-drying, followed by rehumidification
at constant relative humidities (RHs). The equilibrium water contents and Tg values of the sucrose–
additive mixture were then measured. The proteinstabilizing effects of the various sugar–additive
mixtures were evaluated, based on the residual
activity of a model enzyme after freeze-drying. Based
on these experimental results, the impacts of various
additives on the amorphous sucrose matrix were
categorized and possible mechanisms are discussed.
MATERIALS AND METHODS
Materials
Sucrose was purchased from Wako Pure Chemical
Industries, Ltd, (Osaka, Japan). Four dextrans,
DX1.5k (MW 1500 enzymatic synthesis), DX6k
(MW 6000 from Leuconostoc sp.), DX20k (MW
15,000–20,000 from Leuconostoc sp.), and DX500k
(MW 500,000 from Leuconostoc sp.), were products of
Fluka Chemie GmbH (Buchs, CH, Switzerland).
Xanthan (MW 900,000–1,600,000), curdlan (MW
590,000), Pullulan PF-20 (MW 200,000), Pullulan
P400 (380,000) were obtained from Hayashibara
Biochemical Laboratories, Inc., (Okayama, Japan).
Two polyvinylpyrrolidones (PVPs) with molecular
weights of ca. 40,000 (PVP K25) and ca. 1,300,000
(PVP K90), b-cyclodextrin (b-CD), b-cyclodextrin (bCD), sodium glucuronic acid (GA) monohydrate and
glucose-6-phosphate (G6P) disodium salt were from
Wako Pure Chemical Industries, Ltd. Glycerol phosphate (GP) disodium salt hydrate and phytic acid (PA)
sodium salt were products of Sigma-Aldrich Co. (St.
Louis, MO). Hyaluronic acid sodium salt, dipotassium
hydrogenphosphate, and monobasic potassium phosphate were from Nacalai tesque (Kyoto, Japan). P2O5,
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010
LiCl, CH3COOK, and MgCl2 were obtained from Wako
Pure Chemical Industries, Ltd.
Alkaline phosphatase (ALP, from calf intestine) in a
50% glycerol solution was obtained from Toyobo, Co.
(Osaka, Japan) and was used as a model protein. ALP
was used without further purification because it was
diluted by more than 15,000-fold with a buffer
solution to prepare the sample solutions containing
sucrose, additive, and ALP. p-Nitrophenylphosphate
( p-NPP) and diethanolamine were purchased from
Wako Pure Chemical Industries, Ltd.
All other chemicals were of reagent grade.
Preparation of Samples
Amorphous sucrose containing various types and
contents of additives were prepared in the same
manner as described in our previous study.13 In an
typical experiment, an aqueous solution containing
sucrose and additive at a total concentration of 50 mg/
mL was instantaneously frozen with liquid nitrogen
and the frozen sample then freeze-dried at room
temperature for 24 h in a glass chamber, which was
evacuated using a vacuum pump connected to a vapor
condenser.22 The freeze-dried samples were thoroughly dehydrated at 378C by storage in a vacuum
desiccator over P2O5 for more than 2 days and then
rehumidified at 258C in a vacuum desiccator in the
presence of saturated solutions of LiCl, CH3COOK,
and MgCl2 for about 1 week. Preliminary experiments confirmed that the water content reached
equilibrium within a few days, by comparing the
water contents for different humidification periods.
The RH values in the desiccators with saturated
LiCl, CH3COOK, and MgCl2 solutions were 11%,
23%, and 33%, respectively.23
To evaluate and compare the protein-stabilizing
effects of the different sucrose–additive mixtures, the
model protein (ALP) was freeze-dried in the presence
of various sucrose–additive mixtures as follows.
Typically, sucrose and the additive were dissolved
at a total concentration of 50 mg/mL in a 10 mM
diethanolamine buffer (pH 9.5) containing 0.25 mM
MgCl2 instead of distilled water. The stock solution of
ALP (1.2 mg/mL ALP solution containing 50% glycerol) was then diluted with the sucrose–additive
mixture solution to give a final concentration of 72 ng/
mL, and the resulting solution was cooled on an ice
bath. The resulting solution containing ALP was then
immediately freeze-dried in the same manner as
described above. After 24-h of freeze-drying, the
dried samples were reconstituted with distilled
water and residual ALP activities measured, as
described below.
Water Content Analysis
The water contents of the sucrose–additive mixtures
were determined by the differences between weights
DOI 10.1002/jps
PHYSICAL PROPERTIES OF SUCROSE–ADDITIVE MIXTURES
before and after rehumidification in the presence of
saturated salt solutions, as described above. Alternatively, a Karl–Fischer titration was also conducted
in the same manner as in our previous study.6
4671
The findings indicated that, at most, 0.2% of the total
was released into the matrix.
All analyses were performed in triplicate or more
for each sample.
Differential Scanning Calorimetry
The physical stability of an amorphous sugar matrix
is frequently represented by the Tg value. Hence, the
values of Tg for the dried and rehumidified sucrose–
additive mixtures were determined by differential
scanning calorimetry (DSC) using a Perkin-Elmer
DSC Pyris (Norwalk, CT) in the same manner as was
used in our previous study.24 Two to 5 mg samples
were transferred into 20 mL aluminum pans under a
nitrogen atmosphere and hermetically sealed. The
samples were scanned at a rate of 108C/min from a
prescribed temperature (at least 508C lower than the
Tg) to 1808C, using an empty aluminum pan as a
reference. From the obtained thermograms, Tg values
were determined as the onset temperatures of shifts
in apparent specific heat due to transitions.
Alternatively, prewarming to a temperature at
least 108C higher than the Tg was conducted to
eliminate the structural enthalpy relaxation of an
amorphous mixture prior to the whole scanning.25,26
However, no detectable (less than 38C) differences
were found between the two results with and without
the prewarming.
In all the obtained thermograms, each sample
showed a single thermal event due to the glass-torubber transition of the amorphous matrix, indicating
that sucrose and additive molecules were homogeneous in the mixture without any significant phase
separation.
RESULTS AND DISCUSSION
Water Sorption Behavior of Amorphous
Sugar–Additive Mixtures
The equilibrium water contents for amorphous
sucrose containing various types and contents of
additives at constant RHs were measured and the
representative results are shown in Figure 1. The
curves for the relationships between water content
and mixture composition can be roughly classified
into two groups. One is convex downward with
respect to the additive content (Fig. 1a), and the
other is straight between the water contents for
sucrose alone and additive alone or slightly upward
Enzyme Activity Assay
After freeze-drying, the resulting sucrose–additive
mixtures containing ALP were rehydrated with
distilled water to give the same composition as that
before freeze-drying and heated to 378C together with
the reaction mixture containing 1 M diethanolamine,
0.25 mM MgCl2, and 100 mM p-NPP (pH 10.5). A
100 mL aliquot of the rehydrated sample was then
added to 2.9 mL of the reaction mixture. Immediately
thereafter, the increase in absorbance at 405 nm was
measured using a Shimadzu UV-1600 spectrometer
(Shimadzu Co., Kyoto, Japan), from which the initial
reaction rate was determined, in order to evaluate the
extent of preservation of ALP activity by the sucrose–
additive mixture.
It is possible that G6P, GP, and PA could dephosphorylated by ALP before and during the freezedrying, and this could affect the protein stabilizing
effect of the matrix mixture. Hence the concentration
of free phosphate ions, liberated from these compounds by ALP, was determined using AquaCheck P
(Siemens Healthcare-Diagnostics Co., Tokyo, Japan).
DOI 10.1002/jps
Figure 1. Water contents of amorphous sucrose containing (a) dextran 20k and (b) glucose 6-phosphate (G6P). RHs
for rehumidification were 11% (triangles), 23% (circles), and
33% (diamonds).
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010
4672
IMAMURA ET AL.
(Fig. 1b). Mixtures that contained polysaccharides
(dextrans, xanthan, curdlan, pullulan, hyaluronic
acid), cyclodextrins, and PVPs (large-sized additives)
show a downward convex curve. The curves corresponding to phosphate compounds (G6P, GP, PA,
potassium phosphate) and GA belong to the other
group (small-ionized additives) (see Fig. SI1 in
Supplemental Information).
On the other hand, sample solutions containing
more than 33% GP and more than 40% potassium
phosphate dramatically collapsed during freeze-drying (see Fig. SI1 in Supplemental Information). As a
result, the equilibrium water contents of these
samples were not measured and it was not possible
to categorize water sorption behavior of sucrose–GP
and –potassium phosphate mixtures.
As indicated by the data shown in Figure 1, the
relationship between water content and mixture
composition for mixtures containing large molecular
sized additives show downward convex curves. This
suggests that the amount of hydration water for
sucrose and additive in the mixture are lower than
those for sucrose alone and additive alone.13,24,27,28
This can be explained as follows. A larger sized
molecule is likely to form a less dense matrix,28,29
which would contain some dead spaces and thus a
larger number of free side-chain polar groups, such as
hydroxyl groups in polysaccharides, which play a
major participant in the sorption of water. In this
case, when a large sized additive molecule is
embedded in an amorphous matrix comprised of
sucrose, the side-chain polar groups of the additive
molecule would interact with sucrose molecules to a
greater extent than would occur in the absence of the
sucrose matrix by integrating sucrose molecules into
the dead spaces. Consequently, the hydration levels of
large-sized additive molecules in an amorphous
sucrose matrix may be lower than those in the free
states, as suggested in Figure 1a.
On the other hand, when molecules that contain
ionizable groups are embedded in an amorphous
sugar matrix, hydrogen bonds are formed between
the ionizable species and sugar hydroxyl groups.
Hydrogen bonds between ionizable species and sugar
hydroxyl groups may be stronger than sugar–sugar
hydrogen bonds and, hence, would restrict the
allocational and orientational motions of sugar
molecules more tightly. Consequently, the content
of sugar–sugar hydrogen bonds would be decreased
and free hydroxyl groups, which are responsible for
water sorption, would be increased by the presence of
ionizable groups, as shown in Figure 1b.
Hyaluronic acid and xanthan show negative
deviations for water content from ideality, which is
typical for large-sized additives, although these
molecules contain carboxylic groups (see Fig. SI1 in
Supplemental Information). This indicates that the
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010
characteristics of a large molecule are predominant
rather than that of an ionizable group.
As discussed above, the negative deviation in
water content from ideality (Fig. 1a), observed for
mixtures containing large molecular sized additives,
is considered to arise from the formation of hydrogen
bonds between sucrose and additive molecules.
Hence, the amounts of hydration water replaced by
the sucrose–additive hydrogen bonds were estimated
for mixtures containing polysaccharides, CDs, and
PVPs, using the same procedures as was used in our
previous studies.24,27,28 Namely, the water contents of
the mixtures, w (g-water/g-dry matter) were converted to the amount of water involved in a sample
containing 1 g of sucrose, W (g-water/g-sucrose) and
were plotted against the amount of additive added to
1 g of sucrose (Fig. 2). The plots for W were
approximated by one or two lines; The number of
approximated lines indicates the hydration states
and the slopes of the lines are considered to be
proportional to the amount of hydration water for the
additive component.24,28 The water contents of an
additive embedded in amorphous sucrose, w0add (g/gdry additive) were calculated, using the following
equation (1).
w0add ¼
wadd þ dW=dcadd
2
(1)
where wadd (g-water/g-additive) and dW/dcadd (gwater/g-additive) denote the water content for an
additive alone and the slope of the approximated
line for the low additive content range, respectively.
Eq. (1) is based on the assumption that the amount of
Figure 2. Water contents on the basis of sucrose mass for
sucrose–dextran (MW 20k). RHs for rehumidification were
11% (triangles), 23% (circles), and 33% (diamonds). The
lines in figure are the results from the collinear approximation of the plots.
DOI 10.1002/jps
PHYSICAL PROPERTIES OF SUCROSE–ADDITIVE MIXTURES
hydration water replaced from the additive component by sucrose–additive interactions are equal to
those from sucrose molecules.28 Finally, the relative
changes in the amount of hydration water for the
additive components that are embedded in an
amorphous sucrose matrix, Dwadd (%), were obtained
using the following Eq. (2) and are listed in Table 1 as
well as the values for wadd and dW/dcadd.
Dwadd
ðw0 wadd Þ
¼ add
wadd
(2)
As shown in Table 1, the amounts of hydration
water for dextrans, hyaluronic acid, and CDs in an
amorphous sucrose matrix were decreased significantly (13–35%) as the result of being embedded in an
amorphous sucrose. The decrease in the hydration
4673
levels for xanthan and curdlan in an amorphous
sucrose were comparatively small, suggesting that
b-1,3-glycosidic bonds in xanthan and curdlan,
result in polysaccharide conformations that are less
suitable for hydrogen bonding to small sized sugar
molecules.
Glass Transition Temperatures of Amorphous Sucrose
Mixtures
Strong exothermic peaks were observed at around
1608C (data not shown) in thermograms of amorphous sucrose containing G6P and GP, and the
sucrose–G6P and –GP mixtures that had been heated
above 1608C were found to be orange-colored. These
findings demonstrate that chemical reactions took
place in this temperature range and precluded the
Table 1. Relative Changes in the Amount of Hydration Water for Additives as the Result of Being Embedded in Amorphous
Sucrose, Dwadd
Additives
Dextran (1500)
Dextran (6000)
Dextran (20,000)
Dextran (500,000)
Pullulan PF-20
Pullulan P400
a-Cyclodextrin
b-Cyclodextrin
Curdlan
Hyaluronic acid
Xanthan
PVP K25
PVP K90
RH (%)
wadd (g/g-Dry Matter)
dW/dcadd (g/g-Dry Matter)
Dwadd (%)
11
23
33
11
23
33
11
23
33
11
23
33
11
23
33
11
23
33
11
23
33
11
23
33
11
23
33
11
23
33
11
23
33
11
23
33
11
23
33
0.046
0.068
0.085
0.055
0.077
0.100
0.054
0.082
0.101
0.073
0.096
0.123
0.043
0.058
0.074
0.063
0.085
0.101
0.057
0.084
0.103
0.069
0.111
0.152
0.064
0.087
0.112
0.118
0.153
0.179
0.070
0.093
0.118
0.049
0.086
0.122
0.052
0.090
0.130
0.017
0.025
0.044
0.026
0.035
0.035
0.025
0.044
0.041
0.048
0.056
0.064
0.026
0.034
0.055
0.036
0.054
0.063
0.041
0.053
0.053
0.043
0.054
0.047
0.044
0.058
0.076
0.082
0.087
0.096
0.060
0.072
0.089
0.035
0.048
0.060
0.037
0.056
0.076
32
32
24
26
28
32
27
24
30
17
21
24
19
21
13
21
18
19
14
18
24
19
25
35
16
16
16
15
22
23
7
11
12
14
22
25
14
19
21
The values for Dwadd were calculated, using Eq. (1), from the water contents for additives alone (wadd) and initial slopes of the lines for the relationship between
water content a sample containing 1 g of sucrose and the amount of additive added to 1 g of sucrose (dW/dcadd).
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010
4674
IMAMURA ET AL.
measurement of the Tg values for sucrose–G6P and –
GP mixtures that would be expected to be higher than
1308C. Furthermore, the Tg values for amorphous
matrices formed solely by adding PA, b-, b-cyclodextrin, and pullulans also could not be measured
because they were higher than the usable upper
limit temperature of the DSC pan (1808C).
The measured Tg values of the representative
sugar–additive mixtures rehumidified at different
RHs are plotted as a function of additive content and
are shown in Figure 3. Polysaccharides, cyclodextrins, and PVPs clearly show different additive
content dependences for Tg from low-molecularweight additives with ionizable groups (see Fig. SI2
in Supplemental Information), similar to the case for
the relationship between water content and additive
composition, as described above: the addition of largesized additives results in a significant increase in Tg
in the additive content region above 40% (Fig. 3a).
When the RH is 33%, a marked rise of Tg is also
Figure 3. Glass transition temperatures, Tg, of amorphous sucrose (a) dextran 20k and (b) glucose 6-phosphate
(G6P). RHs for rehumidification were 11% (triangles), 23%
(circles), and 33% (diamonds). Results for thoroughly dehydrated samples (RH 0%) are represented by squares.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010
observed in the low additive content region below ca.
25%. On the other hand, in the case of small-ionized
additives (G6P, GP, GA, and PA), the Tg value
increases significantly as the additive content
increases from zero and shows an upward convex
curve with respect to additive content (Fig. 3b).
However, the increase in Tg by the addition of smallionized additive at RHs 11% is much smaller than
that at RH 0%.
In a binary system of sugar and additive, three
types of intermolecular interactions, namely, hydrogen bond formation between sugar molecules,
between sugar and additive, and between additive
molecules, are possible. Sugar–sugar and sugar–
additive hydrogen bonds are dominant at low additive
contents and when the additive content increases
above a certain level, additive–additive interactions
become significant, while sugar–sugar hydrogen
bonds decrease significantly and essentially disappear. Considering these findings, the marked
increase in Tg, observed in the low additive content
range for mixtures containing small-ionized additives
(Fig. 3b), suggest that sugar–additive interactions are
stronger than sugar–sugar hydrogen bonds. This is in
good agreement with the conclusions based on the
water sorption of the sugar–additive mixtures
(Fig. 1b), as described above. On the other hand,
hydrogen bonds between sucrose and large-sized
additives are considered to be approximately equal in
strength, or only slightly stronger than sucrose–
sucrose hydrogen bonds, as indicated by quite the
slight increase in Tg in the low additive content region
at RH 0–23%. Whereas marked increases in Tg in the
additive content range (<25%) at RH 33% were
observed in our previous study13,24 and explained as
follows:13,24 hydration water for an amorphous sugar
matrix serves to weaken the mean strength of
interactions holding the matrix together. When the
strength of the mean interaction for the matrix is
decreased below that for sucrose–additive interactions, sucrose–additive interactions become more
relevant and can contribute to holding the matrix
in the solid phase and the consequent increase in Tg.
The marked increase in Tg in the additive content
range above 40% can be attributed to the formation of
intermolecular interactions between the large molecular sized additives, as indicated by the much higher
Tg values for the additives alone than for sucrose.
As shown in Figure 3b, the sucrose–G6P mixture
exhibited a maximum at around a 50% additive
content when the RH is 33%. The Tg values for
mixtures containing potassium phosphate and GP are
also considered to have a maximum at certain
additive contents although the Tg values for the GP
or potassium phosphate contents above 40% could
not be measured because of the significant collapse
that occurred during freeze-drying, as described
DOI 10.1002/jps
PHYSICAL PROPERTIES OF SUCROSE–ADDITIVE MIXTURES
4675
above (see Fig. SI2 in Supplemental Information). On
the other hand, some equations related to correlating
the Tg values of a binary system with the composition
have been reported.25,30,31,32 Such Tg-composition
relationships having a maximum Tg value cannot be
approximated by the correlation equations proposed
by Fox and Flory,25 Gordon and Taylor,30 and
Couchman and Karasz,31 which assume that two
components that make up the system of interest are
miscible and that the free volumes of the two
components are additive, but the equations proposed
by Kwei32 and the Schneider equations33 that include
the contribution of interactions between the two
components can be used successfully.34 This appears
to support the above consideration regarding the
contribution of sugar–additive interactions to Tg.
The increase in Tg of amorphous sugars by the
presence of phosphate salts has been reported by
Ohtake et al.18 who concluded that this was due to the
fact that phosphate interacts with sugar hydroxyl
groups more tightly than other sugar molecules,
based on FTIR analyses of sugar–phosphate mixtures.18 The marked increase in Tg by the addition of
GP, G6P, and PA also appear to be due to hydrogen
bond formation between phosphate and sucrose
hydroxyl groups. It was found that the Tg values
for mixtures with phosphate compound mixtures
was higher than that for a mixture containing a
phosphate salt at a given additive content (see Fig.
SI2 in Supplemental Information), probably because
the covalent bond to a pyranose ring greatly restricts
the motion of the phosphate, which is responsible for a
glass-to-rubber transition.
The Tg values for amorphous sucrose containing
20% PA was measured at different pHs (Fig. 4). As
shown in Figure 4, an increase in pH tends to result in
an increase in the Tg of a sucrose–PA mixture,
which is in agreement with the reported pH
dependence of Tg for amorphous sugars containing
phosphate salts.18 This indicates that the—PO4Na2
group is a strong contributor to the increase in Tg by
the addition of PA.
The addition of GA increases the Tg value to a
greater extent than xanthan and hyaluronic acid at
RH0% although xanthan and hyaluronic acid also
contain carboxylic groups as well as GA (see Fig. SI2
in Supplemental Information; 89, 70, and 788C at 25%
GA, xanthan, and HA, respectively). This may be
closely related to the distribution of carboxylic
groups, which are probably responsible for the Tg
increase: the carboxylic groups of xanthan and HA
are microscopically localized around the main chains
while those of GA are distributed over the entire
sample at the molecular level. Locally formed
hydrogen bonds between polysaccharide carboxylic
groups and sucrose hydroxyl groups would not serve
to strengthen the amorphous sucrose matrix as
effectively as uniformly distributed hydrogen bonds
would.
The Tg value for an amorphous sugar matrix
decreases with increasing amounts of sorbed water.19
In order to compare the effects of the various additives
on the sensitivity of Tg to the water content, the
average decrease in Tg for the sucrose–additive
mixtures per 0.01 g/g-dry matter of water content,
DTg/w0.01 (8C/[0.01 g-water/g-dry sample]) were
derived from the water contents and Tg values at
different RHs and are shown in Figure 5 (and Fig. SI3
in Supplemental Information). In the case of phosphate salts, PA, and GA, the decrease in Tg with
increasing water content became less significant with
increasing water content. Hence, the DTg/w0.01 values
for amorphous sucrose containing these additives
were determined solely from Tg and water content
data for RHs 0 and 11%.
As shown in Figure 5, the effects of various
additives on DTg/w0.01 values are also classified into
two classes. The additions of polysaccharides, CDs,
and PVPs do not significantly change the sensitivity
of Tg to an increase in water content, as shown in
Figure 4. Tg of amorphous sucrose containing 20 wt.% of
phytic acid at different pHs. RH was 0%.
Figure 5. Changes in Tg of amorphous sucrose–additive
mixtures due to increase in water content by 0.01 g-water/gdry matter, DTg/Dw0.01. Additives were dextran 20k (open
circles) and glucose 6-phosphate (closed circles).
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010
4676
IMAMURA ET AL.
Figure 5. In contrast, in the presence of phosphate
compounds (G6P, GP, PA, potassium phosphate) and
GA, the decrease in Tg due to the increased water
content is greater than that for sucrose alone (Fig. 5).
As described above, the addition of these additives
significantly increases Tg in the thoroughly dehydrated state (RH 0%) (Fig. 3b) while it increases the
amount of sorbed water (Fig. 1b). Consequently, the
effect of small-ionized additive on significantly
increasing Tg would be somewhat compensated by
the increase in the amount of sorbed water, resulting
in the increase in the sensitivity of Tg to the water
content in the rehumidified state (RHs 11%, 23%, and
33%). This may be the reason for why the increase in
Tg values as the result of the addition of small-ionized
additives is much smaller at RHs 11% than at
RH 0%.
Stabilizing Effect of Sugar–Additive Mixtures on Enzyme
Activity during Freeze-Drying
Figure 6 shows residual ALP activities after freezedrying in the presence of various combinations of
sucrose and additives as well as for sucrose alone. In
the absence of any stabilizing agent, the ALP activity
was decreased below the limit of detection, and when
a 50-mg/mL solution of sucrose was solely added to
the ALP solution, about 94% of the ALP activity was
maintained after freeze-drying. Alternatively, the
remaining ALP activities after freeze-drying with
different concentrations of sucrose were also measured and found to reach a plateau at the sucrose
content 10 mg/mL. The ALP concentration also did
not affect the remaining ALP activity in the tested
range (10–200 ng/mL). Hence, the remaining ALP
activities shown in Figure 6 are considered to be
the maximum values respectively for sucrose and
probably various compositions of sucrose–additive
mixtures.
Mixtures containing dextran 6k, phosphate salts,
G6P, and GP also show an ALP activity approxi-
Figure 6. Remaining ALP activities in different sucrose–
additive mixtures after freeze-drying. Deviations in relative
remaining ALP activities for more than three measurements were within 7% of the average.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010
mately equal to or slightly lower than the sucrose
alone sample in the low additive content range below
25 wt.%. The ALP activity for sucrose–PA and –GA
mixtures were decreased by 20% during freezedrying at additive contents of 10–25%. The addition of
dextran 20k and b-CD resulted in a marked decrease
in residual ALP activity. When the additive content
was increased above that value, the residual ALP
activity was dramatically decreased, except for the
case of GA. It is noteworthy that the addition of more
than 50 wt.% of PA completely abolishes ALP activity
during freeze-drying.
Considering these experimental results, the protein-stabilization effect and the physical stability of a
sugar–additive mixture generally appears to be in a
dilemmatic relationship: a sugar–additive mixture
having a higher Tg tends to show a lower residual
ALP activity after freeze-drying. On the other hand,
phosphate salts and GA may be the most effective of
the tested additives because ALP activity was
maintained at relatively constant levels, even at
comparatively high additive contents. The addition of
small amounts of dextran 6k may be effective for use
as an additive component under humid conditions, as
opposed to K-phosphate and GA, because mixtures
containing phosphate salts and GA show a marked
decrease in Tg due to water sorption, as shown in
Figure 5.
CONCLUSION
One possible strategy for improving both the proteinstabilizing effect and physical stability of an amorphous sugar matrix is to combine disaccharides,
which are responsible for protein stabilization, with
an additive, which functions to hold the matrix in
place.5,6 In this study, amorphous sucrose matrices
containing various types and contents of additive
substances were prepared by freeze-drying and the
effects of the additives on water sorption behavior,
glass-to-rubber transition temperatures, and protein
stabilization during freeze-drying were investigated.
The additives tested could be classified into two
groups by their effect on water sorption and Tg, and
both groups of additives had both merits and
demerits: the addition of polysaccharides, cyclodextrins, or PVPs gave only a relatively small increase in
Tg in the low additive content range but decreased the
water sorption. In contrast, the addition of small
amounts of a low MW substance having ionizable
groups resulted in a marked increase in Tg and,
however, enhanced the lowering in Tg due to water
sorption. The protein-stabilizing effect of the sucrose–
additive mixture was as high as or slightly lower than
that of a sucrose alone matrix when the additive
content was low (<25 wt.%) but was significantly
DOI 10.1002/jps
PHYSICAL PROPERTIES OF SUCROSE–ADDITIVE MIXTURES
decreased with an increase in additive content above
certain levels.
15.
ACKNOWLEDGMENTS
16.
This work was supported by grant-in-aids for the
Encouragement of Young Scientists (no. 18760594)
and Science Research (C) (no. 20560702) from the
Ministry of Education, Science, Sport and Culture of
Japan, the Iijima Foundation for Food Science, and the
Okayama Foundation for Science and Technology.
REFERENCES
1. Pikal MJ. 1994. Freeze-drying of proteins. Process, formulation
and stability. In: Clel JL, Langer R, editors. Formulation and
delivery of proteins and peptides, ACS Symposium Series,
Washington DC: American Chemical Society, Vol. 567,
pp 120–133.
2. Carpenter JF, Prestrelski SJ, Anchodorquy TJ, Arakawa T.
1994. Interactions of proteins with stabilizers during freezing
and drying. In: Clel JL, Langer R, editors. Formulation and
delivery of proteins and peptides, ACS Symposium Series,
Washington DC: American Chemical Society, Vol. 567,
pp 134–147.
3. Franks F. 1991. Long-term stabilization of biologicals. Biotechnology 12:38–42.
4. Prestrelski SJ, Pikal KA, Arakawa T. 1995. Optimization of
lyophilization conditions for recombinant human interleukin-2
by dried-state conformational analysis using Fourier-transform infrared spectroscopy. Pharm Res 12:1250–1259.
5. Allison SD, Manning MC, Randolph TW, Middleton K, Davis A,
Carpenter JF. 2000. Optimization of storage stability of lyophilized actin using combinations of disaccharides and dextran.
J Pharm Sci 89:199–214.
6. Imamura K, Ogawa T, Sakiyama T, Nakanishi K. 2003. Effect
of types of sugar on the stabilization of protein in the dried
state. J Pharm Sci 92:266–274.
7. Slade L, Levine H. 1988. Non-equilibrium behavior of small
carbohydrate-water system. Pure Appl Chem 60:1841–1864.
8. Izutsu K, Yoshioka S, Takeda Y. 1991. The effect of additives on
the stability of freeze-dried b-galactosidase stored at elevated
temperature. Int J Pharm 71:137–146.
9. Suzuki T, Imamura K, Fujimoto H, Okazaki M. 1998. Relation between thermal stabilizing effect of sucrose on LDH
and sucrose-LDH hydrogen bond. J Chem Eng Japan 31:
565–570.
10. Zhang Y, Ji B, Ling P, Zhang T. 2007. Trehalose and hyaluronic
acid coordinately stabilized freeze-dried pancreatic kininogenase. Eur J Pharm Biopharm 65:18–25.
11. Orford PD, Parker R, Ring SG. 1990. Aspects of the glass
transition behavior of mixtures of carbohydrates of low molecular weight. Carbohydr Res 196:11–18.
12. Taylor LS, Zografi G. 1998. Sugar-polymer hydrogen bond
interactions in lyophilized amorphous mixtures. J Pharm Sci
87:1615–1621.
13. Imamura K, Fukushima A, Sakaura K, Sugita T, Sakiyama T,
Nakanishi K. 2002. Water sorption and glass transition behaviors of freeze-dried sucrose-dextran mixtures. J Pharm Sci
91:2175–2181.
14. Imamura K, Ohyama K, Yokoyama T, Maruyama Y, Imanaka
H, Nakanishi K. 2008. Temperature scanning FTIR analysis of
DOI 10.1002/jps
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
4677
interactions between sugar and polymer additive in amorphous
sugar-polymer mixture. J Pharm Sci 97:519–528.
Miller DP, de Pablo JJ, Corti HR. 1999. Viscosity and glass
transition temperature of aqueous mixtures of trehalose
with borax and sodium chloride. J Phys Chem B 103:
10243–10249.
Kets EPW, Ijpelaar PJ, Hoekstra FA, Vromans H. 2004. Citrate
increases glass transition temperature of vitrified sucrose preparations. Cryobiol 48:46–54.
Wolkers WF, Oldenhof H, Tablin F, Crowe JH. 2004. Preservation of dried liposomes in the presence of sugar and phosphate.
Biochim Biophys Acta 1661:125–134.
Ohtake S, Schebor C, Palecek SP, de Pablo JJ. 2004. Effect
of pH, counter ion, and phosphate concentration on the glass
transition temperature of freeze-dried sugar-phosphate mixtures. Pharm Res 21:1615–1621.
Roos Y, Karel M. 1991. Plasticizing effect of water on thermal
behavior and crystallization of amorphous food model. J Food
Sci 56:38–43.
Bakhit R, Schmidt S. 1993. Sorption behavior of mechanically
mixed and freeze-dried sucrose/casein mixture. J Food Sci
58:1162–1165.
Chinachoti P, Steinberg MP. 1988. Interaction of sucrose with
gelatin, egg albumin and gluten in freeze-dried mixtures as
shown by water sorption. J Food Sci 53:932–934.
Imamura K, Maruyama Y, Tanaka K, Yokoyama T, Imanaka
H, Nakanishi K. 2008. True density analysis of a freeze-dried
amorphous sugar matrix. J Pharm Sci 97:2789–2797.
Greenspan L. 1977. Humidity fixed points of binary saturated
aqueous solutions. J Res Nat Bur Stand Sect A 81:89.
Imamura K, Asano Y, Maruyama Y, Yokoyama T, Nomura M,
Ogawa S, Nakanishi K. 2008. Characteristics of hydrogen
bond formation between sugar and polymer in freeze-dried
mixtures under different rehumidification conditions and its
impact on the glass transition temperature. J Pharm Sci 97:
1301–1312.
Fox TG, Flory PJ. 1950. Second-order transition temperatures
and related properties of polystyrene. I. Influence of molecular
weight. J Appl Phys 21: 581–591.
Tant MR, Wilkes GL. 1981. An overview of the nonequilibrium
behavior of polymer glasses. Polym Eng Sci 21:874–895.
Costantino HR, Curley JG, Wu S, Hsu CC. 1998. Water sorption behavior of lyophilized protein-sugar systems and implications for solid-state interactions. Int J Pharm 166: 211–221.
Imamura K, Iwai M, Ogawa T, Sakiyama T, Nakanishi K. 2001.
Evaluation of hydration states of protein in freeze-dried amorphous sugar matrix. J Pharm Sci 90:1955–1963.
van den Dries IJ, van Dusschoten D, Hemminga MA, van der
Linden E. 2000. Effects of water content and molecular weight
on spin probe and water mobility in malto-oligomer glasses.
J Phys Chem B 104:10126–10132.
Gordon M, Taylor JS. 1952. Ideal co-polymers and the second
order transitions of synthetic rubbers. 1. Non-crystalline copolymers. J Appl Chem 2:493–500.
Couchman PR, Karasz FE. 1978. A classical thermodynamic
discussion of the effect of composition on glass transition
temperatures. Macromolecules 11:117–119.
Kwei TK. 1984. The effect of hydrogen bonding on the glass
transition temperatures of polymer mixtures. J Polym Sci Pol
Lett 22:307–313.
Schneider HA. 1989. Glass transition behavior of compatible
polymer blends. Polymer 30:771–779.
Tajber L, Corrigan OI, Healy AM. 2005. Physicochemical evaluation of PVP-thiazide diuretic interactions in co-spray-dried
composites-analysis of glass transition composition relationships. Eur J Pharm Sci 24:553–563.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010