Embed Size (px)
Citation preview
1
New scale-down methodology from commercial to lab scale to optimize
plant-derived soft gel capsule formulations on a commercial scale
Sana Oishia, Shin-ichiro Kimura
a, Shuji Noguchi
b, Mio Kondo
c, Yosuke Kondo
c, Yoshiyuki 5
Shimokawac, Yasunori Iwao
a,*, Shigeru Itai
a,*
aSchool of Pharmaceutical Sciences, University of Shizuoka, 52-1 Yada, Suruga-ku, Shizuoka
422-8526, Japan 10
bFaculty of Pharmaceutical Sciences, Toho University, 2-2-1 Miyama, Funabashi, Chiba 274-8510,
Japan
cFuji Capsule Co., Ltd., 4242-1 Kitayama, Fujinomiya, Shizuoka 418-0112, Japan
*To whom correspondence should be addressed. 15
Yasunori Iwao, Ph.D.
Tel.: +81 54 264 5612; fax: +81 54 264 5615; E-mail address: [email protected]
Shigeru Itai, Ph.D.
Tel.: +81 54 264 5614; fax: +81 54 264 5615; E-mail address: [email protected].
20
2
Graphical abstract
25
30
35
40
3
Abstract
A new scale-down methodology from commercial rotary die scale to laboratory scale was developed 45
to optimize a plant-derived soft gel capsule formulation and eventually manufacture superior soft gel
capsules on a commercial scale, in order to reduce the time and cost for formulation development.
Animal-derived and plant-derived soft gel film sheets were prepared using an applicator on a
laboratory scale and their physicochemical properties, such as tensile strength, Young’s modulus, and
adhesive strength, were evaluated. The tensile strength of the animal-derived and plant-derived soft 50
gel film sheets was 11.7 MPa and 4.41 MPa, respectively. The Young’s modulus of the
animal-derived and plant-derived soft gel film sheets was 169 MPa and 17.8 MPa, respectively, and
both sheets showed a similar adhesion strength of approximately 4.5–10 MPa. Using a D-optimal
mixture design, plant-derived soft gel film sheets were prepared and optimized by varying their
composition, including variations in the mass of κ-carrageenan, ι-carrageenan, oxidized starch and 55
heat-treated starch. The physicochemical properties of the sheets were evaluated to determine the
optimal formulation. Finally, plant-derived soft gel capsules were manufactured using the rotary die
method and the prepared soft gel capsules showed equivalent or superior physical properties
compared with pre-existing soft gel capsules. Therefore, we successfully developed a new
scale-down methodology to optimize the formulation of plant-derived soft gel capsules on a 60
commercial scale.
Key words: Soft gel capsule; scale-down methodology; rotary die method; design of experiment;
gelatin; carrageenan.
65
Abbreviations: DoE, design of experiment; MCT, medium chain triglyceride.
4
1. Introduction 70
Soft gel capsules are dosage forms in which active ingredients can be enclosed with oily
liquids, semi-liquids and pastes (Stanley, 1986; Shah et al., 1992), and are now used in various fields
such as pharmaceuticals (Bottom et al., 1997; Cuppoletti et al., 2004), foods, and health foods
(Jannin et al., 2015; Maki et al., 2012). Soft gel capsules can mask unpleasant odors and tastes
derived from the active ingredients contained within the capsule. They have excellent airtightness, 75
provide stability for their filled contents, and rapidly disintegrate in the stomach when they are orally
administered (Gullapalli, 2010). To date, almost all drug candidates developed by the pharmaceutical
industry are known to show poor water solubility (Lipinski, 2002), and soft gel capsules have
attracted attention because they can encapsulate oily poorly water-soluble drugs that have been
solubilized inside the capsules. 80
There are two ways to manufacture soft gel capsules; one is the rotary die method (Misic et
al., 2012) and the other is the falling-drop method (Homar et al., 2007). In both manufacturing
processes, the capsule shells are usually formed from gelatin. The gelatin is heated to a sol state and
then molded to form shells by means of a sol-gel transition. Of these, the rotary die method is
considered to be suitable for mass production, and many soft gel capsule companies adopt this 85
method (Gullapalli, 2010). A rotary die machine has a pair of roller dies and this method involves the
formation of two gelatin sheets through the use of a gravity-fed spreader box. The technique
involves cooling the liquid gelatin on two separate roller dies, then lubricating and guiding the two
sheets into contact with each other between two co-acting dies while simultaneously dispensing the
proper amount of filling material between the sheets in half-cavities in the outer surface of the dies 90
(Gullapalli, 2010) Therefore, when manufacturing soft gel capsules, the strength and adhesive force
of the sheets are of particular importance to keep the filling materials encapsulated for a long time
(Kamiya et al., 2014), and the flexibility of the sheets in withstanding force from external impacts is
also an essential physical property.
Gelatin has been used for a long time as the main ingredient of the shell for soft gel capsules 95
(Reich, 2004). Gelatin can be obtained from collagen derived from pig skins or cow skins and bones
(Morrison et al., 1999; Schrieber and Gareis, 2007). Since gelatin has high strength and adhesive
strength, it was found to be suitable for soft gel capsule molding (Tesconi et al., 1999). However,
animal gelatin has some disadvantages, such as a specific odor and limitations in terms of its broad
applicability because of the religious or dietary customs of some patients (Badii and Howell, 2006; 100
Misic et al., 2012). In addition, its physicochemical characteristics, such as low stability in response
to increased heat (Nazzal and Wang, 2001), and brittleness because of its high hardness are also of
concern. For these reasons, a search for non-animal shell materials, such as a plant-derived material
alternative to gelatin, has been conducted by academic and industrial researchers (Gullapalli and
Mazzitelli, 2017). 105
5
Recently, we successfully developed plant-derived soft gel capsules, using carrageenan and
starches as the main shell agent ingredients, using the rotary die method (Kato et al., 2013). The
plant-derived soft gel capsules prepared using these materials were found to be modestly soft,
resistant to cracking and stable on heating. However, this formulation (the formulation of
plant-derived soft gel capsule was shown in Table 1) required optimization for further improvement 110
of the physical properties of plant-derived soft gel capsules.
In general, when optimizing a formulation, statistical analysis using a combination of design
of experiment (DoE) and multiple regression analysis is performed using laboratory-scale
experimental equipment and conditions. However, if the formulation optimization of this
plant-derived soft gel capsule was conducted using industrial machines, rotary dies, an enormous 115
investment of time and cost would be required. In addition, neither an evaluation method for the
physicochemical properties of capsule sheets nor their evaluation criteria at a laboratory-scale has
been available for soft gel capsules. Therefore, an evaluation method at a laboratory-scale which can
sufficiently reflect the relationship between the physical properties of capsule sheets and soft gel
capsules is required. The establishment of a scale-down method for manufacturing superior soft gel 120
capsules is imperative and this would be extremely useful when developing various new
plant-derived formulation substitutes, with a view to scaling up in the future.
In terms of this type of evaluation at a laboratory-scale, only one study has been reported in
which film sheets imitating soft gelatin capsule shells were prepared using an applicator for thin
layer chromatography, and the strength of the film sheets was evaluated (Kamiya et al., 2014). In the 125
present study, firstly, soft gel film sheets were prepared using an applicator, based on gelatin and
plant-derived soft gel capsule formulations often used in the rotary die method. Secondly, the
physical properties of the film sheets were determined by evaluating their strength, flexibility and
adhesion. Thirdly, a D-optimal mixture design was selected as the experimental design, and we
attempted to optimize the formulation of the plant-derived soft gel capsule shells using a multiple 130
regression analysis. Finally, plant-derived soft gel capsules were prepared using the rotary die
method based on the optimum plant-derived shell formulation, and their characteristics were
evaluated in detail and compared with pre-existing plant-derived soft gel capsules.
6
2. Materials and Methods 135
2.1 Materials
Gelatin derived from pig skins and bones, which had undergone an alkali-treatment, was
purchased from Nippi Co., Ltd. (Tokyo, Japan). κ-carrageenan (NEWGELIN CT-1000) and
ι-carrageenan including 20% sucrose (MSC10018) were purchased from Mitsubishi Shoji Foodtech
Co., Ltd. (Tokyo, Japan) and MSC Co., Ltd. (Gyeongram, Korea), respectively. Oxidized starch 140
(from potato and tapioca) and heat-treated starch (from corn) were purchased from Matsutani
Chemical Industry Co., Ltd. (Hyogo, Japan) and Sanwa Starch Co., Ltd. (Nara, Japan), respectively.
Food additive grade glycerin and glycerin fatty acid ester (POEM FB-28) were purchased from
Sakamoto Yakuhin Kogyo Co., Ltd. (Osaka, Japan) and Riken Vitamin Co., Ltd. (Tokyo, Japan),
respectively. 145
2.2 Preparation of film sheets for the soft gel capsules
The gelatin and plant-derived soft gel film sheets were prepared following the
formulations shown in Table 1. For the gelatin soft gel film sheets, 100 g of gelatin was added to
100 g of water and 40 g of glycerin in a 500 mL beaker and dispersed. This mixture was heated for 150
30 min with stirring in a water bath at 90°C for 15 min, until the gelatin went into solution. For the
plant-derived soft gel film sheets, water and glycerin and plant-derived materials such as
κ-carrageenan, ι-carrageenan, oxidized starch and heat-treated starch were also mixed in a 500 mL
beaker. Although initially lumpy, the materials went into solution through crushing with stirring for
30 min. Both formulations were then agitated twice every 15 min and then agitated twice every 30 155
min at 95°C. The sol solutions were thoroughly mixed during a total of 2 h of heating. Because both
solutions contained a lot of bubbles during heating, a small amount of glycerin fatty acid ester was
added for degassing and the mixtures were incubated at 90°C for a further 2 h until the bubbles had
completely disappeared. The beakers and the water baths were covered with aluminum foil to
prevent the evaporation of moisture from the solutions. Finally, some hot water was added to the 160
solutions to keep the volume of water in the formulation constant, at approximately 42% and 52%
for the gelatin and plant-derived capsule sheets, respectively.
After preparing homogeneous gelatin and plant-derived mixtures, film sheets were
obtained using an applicator for thin layer chromatography (Advantec Toyo Kaisha, Ltd., Tokyo,
Japan). Briefly, two glass plates and one partition plate were heated to 70°C using a hot plate (AS 165
ONE Corp. Osaka, Japan) and the other partition plate was heated to 100°C using a water bath. The
height of the partition of the applicator was set to 1 mm using a thickness gauge. The solution was
then poured between the partition plates, and the glass plate was quickly slid and allowed to stand
for 2 min. A 2 min incubation was found to be equivalent to when the film sheet was prepared using
the rotary die method, and after 2 min, the solution changed from sol to gel. A uniform film sheet 170
7
with a thickness of approximately 1 mm was obtained.
2.3 Evaluation of film sheets
2.3.1 Tensile strength and flexibility
Tensile testing apparatus (Force Tester®, A&D Co., Tokyo, Japan) was used to examine the 175
strength and flexibility of the film sheets. The film sheets were cut into 60 mm × 20 mm portions
and left to stand at 20°C and 30 ± 5% relative humidity (RH) for 24 h. The length and thickness of
the samples after drying were measured using a digital caliper (CD-15C, Mitutoyo Corp., Kanagawa,
Japan). The dried film samples were then analyzed using a Force Tester® at a test speed of 100
mm/min. The film strength was evaluated using tensile strength at break, and flexibility was 180
evaluated using Young’s modulus, which represents elongation in the elastic range and the
proportional coefficient of stress.
2.3.2 Adhesive strength
In order to evaluate the adhesive strength between the film sheets, they were cut into 185
portions of 30 mm × 30 mm in size. Then, two film sheets were adhered using a film plate type mold
of a simple capsule device and these adhered films were expanded with the adhered portion as the
axis and incubated at 20°C and 30 ± 5% RH for 24 h to completely dry. The dried samples were cut
into 40 mm × 5 mm portions and completely covered with a cellulose tape other than at the adhesion,
to ensure breakage at the bonded part. The samples were assessed using the Force Tester®
at a test 190
speed of 100 mm/min.
2.4 Design of experiment (DoE)
When preparing the film sheets, the amount of glycerin and purified water was maintained
at a constant ratio. The plant-derived materials were chosen as design variables and mixture design 195
was adapted with a constant total mass of 100 g. Four factors were selected; the amount of
κ-carrageenan (X1), ι-carrageenan (X2), oxidized starch (X3) and heat-treated starch (X4). The levels
of these variables were determined by conducting preliminary experiments to determine the range of
factors that can produce film sheets. Especially, the level of κ-carrageenan was set to be 0–3 g
because over 3 g of κ-carrageenan made the films with low strength and no elasticity. The range for 200
ι-carrageenan was from 20 to 35 g because less than 20 g of ι-carrageenan did not give enough
elasticity and flexibility. In addition, over 35 g of ι-carrageenan made the film preparation difficulty
and it made films rough. The determined high and low levels are shown in Table 2. The measured
film properties were tensile strength (Y1), Young’s modulus (flexibility, Y2), and adhesive strength
(Y3) as responses. 205
In the mixture design, other factors such as the manufacturing conditions are considered
8
constant not to influence the response, meaning the responses are functions of the composition in
formulation. The design was performed in accordance with “Experimental with Mixtures” (Scheffe,
1958; Scheffé, 1963) proposed by H. Scheffe, and involved setting the ratio value of the blended
components to 0 or more and 1 or less, and the sum of ratios of them to 1 (Kagamiyama et al., 2003). 210
However, since the setting range of the ingredients is different between the four factors, as shown in
Table 2, it is expected that the constraint conditions and the DoE are complicated. Therefore, it was
shown that the D-optimal design devised by Kiefer was suitable (Kiefer, 1992) as the DoE for
optimizing in such a case.
In other words, after preparing a DoE with a mixture design, the prescription was 215
optimized using a D-optimal mixture design calculated by multiplying the D-optimal design to
optimize the DoE. Design-Expert®
version 10 (Stat-Ease, Inc., Minneapolis, USA) software was
used, which can perform a multivariate analysis and an experimental design. JMP 9® (SAS Institute
Japan Ltd., Tokyo, Japan) was used for optimization of the formulation. The criteria of the three
responses were set based on three items of the existing plant-derived soft gel film formulation that 220
can become soft gel capsules often used in the rotary die method.
2.5 Production of soft gel capsules using the rotary die method
Plant-derived soft gel capsules of an optimized formulation, as assessed using data from
the DoE and optimization study, were manufactured using the rotary die method. The mold for the 225
soft gel capsules was the OVAL 4, which has an elliptical shape with a major diameter of 11 mm and
a minor diameter of 7 mm, and soft gel capsules contain 200 mg of medium chain triglyceride
(MCT). MCT was selected as the model content inside the soft gel capsules.
2.6 Evaluation of the soft gel capsules 230
2.6.1 Dimensional test
The major and minor axis of the soft gel capsules was measured using a digital caliper
(CD-15C, Mitutoyo Corp., Kanagawa, Japan).
2.6.2 Thickness of the sheets and adhesion portion 235
The soft gel capsules were incised, and they were observed using a digital microscope
(VHX-2000, KEYENCE Corp., Osaka, Japan). The thickness of the sheet and the adhesion portion
was measured.
2.6.3 Mass deviation test 240
In accordance with the mass deviation test of The Japanese Pharmacopoeia, Seventeenth
Edition (JP17), the total mass of the soft gel capsules was measured using semi-micro analytical
9
balances (GR-200, A&D Co.). Thereafter, the soft gel capsule was collapsed, to wash out the
encapsulated MCT using ethanol, and left at room temperature for approximately 30 min to
completely evaporate the remaining ethanol. Then, the mass of the empty soft gel capsule was 245
precisely assessed, and the mass of the capsule contents was calculated by subtracting the mass of
the empty capsule from the total mass.
2.6.4 Hardness test
The hardness of the soft gel capsules was measured using a grain rigidity tester (Fujiwara 250
Scientific Co., Ltd., Tokyo, Japan). The soft gel capsules were placed on the tester so that the
adhesive portion was level with the floor surface.
2.6.5 Disintegration test
In accordance with the JP17 disintegration test, a disintegration tester (NT-200, Toyama 255
Sangyo Co., Ltd., Osaka, Japan) was used to measure the time it took for the soft gel capsules to
collapse and completely disintegrate in water (37 ± 2°C).
2.7 Statistics
Statistical analysis was carried out using the F-test and analysis of variance (ANOVA). A 260
probability value of p<0.05 was considered to indicate a significant difference between mean values.
10
3. Results and Discussion
3.1. Evaluation of the film sheets
Standard values for the physical properties of the film sheets were determined by 265
evaluating their strength, flexibility and adhesion.
3.1.1 Tensile strength
When soft gel capsules are produced, if the strength of the film sheet is low, there is
concern that the film sheet could split during the rotary die method and it is difficult to mold the 270
capsule (Kamiya et al., 2014). Therefore, a certain minimum tensile strength is required to mold
capsules. However, when the tensile strength is too high with poor flexibility, it is presumed that the
shell tends to be easily broken. Therefore, it is assumed that there is a range of appropriate tensile
strengths. The tensile strength measured using a tensile tester is the maximum value among the
stresses when the film sheet is stretched in a certain direction. Tensile strength is useful as an 275
indicator of the required strength when soft gel capsules are produced or after being manufactured
into a product.
The tensile strength of the film sheets is shown in Fig. 1 (a). The tensile strengths for the
gelatin and plant-derived soft gel film sheets were found to be 11.7 MPa and 4.41 MPa, respectively.
However, stable manufacturing of soft gel capsules can be performed using both formulations 280
through the rotary die method, indicating that a tensile strength of 4.41 MPa or more is sufficient
when preparing soft gel capsules. Because the tensile strength of the gelatin film sheet was higher
than that of the plant-derived soft gel film sheet, the hardness and crackability of the soft gelatin
capsule described above may be caused by its high strength. This may be influenced by insufficient
elongation to stress within the elastic range. Therefore, the Young’s modulus was evaluated as an 285
index of deformation and flexibility.
3.1.2 Flexibility
The Young’s modulus values of the gelatin and plant-derived soft gel film sheets are
shown in Fig. 1 (b). Young’s modulus is defined as the proportional coefficient obtained by dividing 290
the tensile stress by the strain in the elastic range when the sample is stretched. The lower the value
of the Young’s modulus, the easier it is to deform a material, and it can be described as more flexible.
The Young’s modulus values for the gelatin and plant-derived soft gel film sheets were found to be
169 MPa and 17.8 MPa, respectively.
295
3.1.3 Adhesive strength
The adhesive strength of the gelatin and plant-derived soft gel film sheets is shown in Fig.
1 (c), both sheets showed an adhesion strength of approximately 4.5–10 MPa (6.99 MPa for gelatin
11
soft gel film sheet and 4.66 MPa for plant-derived soft gel film sheet). It is considered that the most
suitable timing to adhere film sheets is immediately after gelation when the shell solution is in a sol 300
state. In our evaluation at the laboratory scale, the film sheets were adhered 2 min after the sheet
preparation, which can be regarded as equivalent timing as that used in the rotary die method. If
adhesion of the film is not performed adequately, it is predicted that cracks could be generated at the
adhesive portion during the manufacturing process or after formation of the capsules. However, both
shell formulations were produced stably and had adequate adhesion. Taken together, it was found 305
that an adhesion strength of 4.66 MPa or more would be suitable on a laboratory scale.
Fig. 1. The physical properties of the film sheets based on the pre-existing formulations (a:
tensile strength, b: Young’s modulus, and c: adhesive strength).
310
12
3.2 Statistical analysis
3.2.1 Design of experiment (DoE)
The experimental designs and responses, expressed as Y1: tensile strength (MPa), Y2:
Young’s modulus (MPa), and Y3: adhesive strength (MPa), are shown in Table 3. A total of 17
experiments were conducted including the center point repeated three times. The results of a multiple 315
regression analysis using the results in Table 3 and the significance of the model as evaluated using
ANOVA are shown in Tables 4–6. In the multiple regression analysis, four independent factors are
called a linear mixture and are expressed as X1, X2, X3, X4, and the effect of the combination of
factors on the film physical properties is referred to as an interaction and expressed as X1X2, X1X3, …,
X3X4. The regression equations (Eqs. (1) – (3)) were obtained through multiple regression analysis of 320
the data for each response. The determination of the statistically significant factors was made based
on the p-value calculated from the F-test.
Y1 = - 350X1 + 67.0X2 + 5.05X3 + 1.88X4 + 142X1X2 + 442X1X3 + 475X1X4 - 79.7X2X3 - 66.8X2X4 -
2.35X3X4 (1) 325
Y2 = - 5147X1 + 487X2 + 48.6X3 + 2.59X4 + 3057X1X2 + 5949X1X3 + 6317X1X4 - 688X2X3 - 544X2X4 -
31.6X3X4 (2)
Y3 = - 1019X1 + 95.9X2 + 5.57X3 + 3.58X4 - 632X1X2 + 1070X1X3 + 1113X1X4 – 117X2X3 - 102X2X4 - 330
6.68X3X4 + 2045X1X2X3 + 1788X1X2X4 + 81.1X1X3X4 + 56.7X2X3X4 (3)
A 3D response surface curve drawn using Design-Expert® version 10 based on a multiple
regression equation (tensile strength, Young’s modulus, and adhesive strength) is shown in Figs. 2–4.
κ-carrageenan (X1) was estimated to be the lowest contribution rate based on the high p-value of the 335
interaction in Tables 4–6 and the original low content such as maximum level of 3.0 g, and this was
fixed to 1.5 g which is the center of the level.
The regression coefficients and p-values of the tensile strength obtained from a multiple
regression analysis are shown in Table 4 and Eq. (1). The model’s coefficient of determination (R2)
was 0.91, and the degree of freedom adjusted coefficient (Rf2) was 0.80, indicating that these values 340
were relatively high and this equation might be valid. The P-value of linear mixture was 0.0012 and
the linear contribution of each factor was found to be large, meaning that it is possible to explain the
tensile strength only by l linear contribution. However, in order to examine this in more detail, a
response surface plot including terms of interaction was created and is shown in Fig. 2. From this, as
indicated by yellow color, it was confirmed that when the oxidized starch (X3) content was high and 345
the heat-treated starch (X4) content was low, the tensile strength increased.
13
Fig. 2. Response surface plot of the mean tensile strength with a fixed amount of κ-carrageenan
(X1) at 1.5 g. 350
In terms of the Young’s modulus, it was also confirmed from the results shown in Table 5
and Eq. (2) that the R2 value of the model was 0.93 and the Rf
2 value was as high as 0.85. From the
results of the response surface plot shown in Fig. 3, as indicated by sky blue, the oxidized starch (X3)
content was low and the heat-treated starch (X4) content was high, and the Young’s modulus 355
decreased and the flexibility was improved. From the tensile strength and Young’s modulus results, it
was shown that a hard gel was formed when a high amount of oxidized starch (X3) was present, and
when the amount of heat-treated starch (X4) was high, the gel seemed to be fragile.
14
Fig. 3. Response surface plot of the mean Young’s modulus with a fixed amount of 360
κ-carrageenan (X1) at 1.5 g.
As for the adhesive strength, it was also confirmed from the results shown in Table 6 and
Eq. (3) that the R2 value of the model was 0.97 and the Rf
2 value was as high as 0.82. From the
results of the response surface plot shown in Fig. 4, it was confirmed that the adhesive strength 365
increased with an increase of the ι-carrageenan (X2) content.
Fig. 4. Response surface plot of the mean adhesive strength with a fixed amount of
κ-carrageenan (X1) at 1.5 g. 370
15
From the DoE results, it was found that the properties of each factor may be strongly
related to each response in terms of the physical properties of the film sheet. κ-carrageenan gave
strength to the film sheet and ι-carrageenan gave elasticity and flexibility. κ-carrageenan has only
one sulphate group per disaccharide repeating unit, whereas ι-carrageenan has two (Liu et al., 2015). 375
These differences in structure were found to give a difference to the gel formation (Piculell et al.,
1992). Specifically, it was found that κ-carrageenan can form hard and brittle gels and ι-carrageenan
can form soft and elastic gels (Torres et al., 2016). Therefore, the number of sulfate groups and the
content of anhydro bridges of carrageenans would be involved in the film properties. From the
results of the 3D response surface curve, it was shown that oxidized starch gave strength to the film 380
sheet and heat-treated starch gave flexibility to the film sheet. Further study using NMR would be
needed to make clear the interaction between starch and carrageenan.
3.2.2 Optimizing the film sheet formulation of the plant-derived soft gel capsule
Based on the results so far, a design space was constructed and is shown in Fig. 5. The 385
values of the physical properties of the existing plant-derived formulation obtained in 3.1 were used
as criteria as shown in Table 7.
The design space when ι-carrageenan (X2) is fixed at 35 g and heat-treated starch (X4) is
fixed at 50 g is shown in Fig. 5. The white region in the design space is the optimal formulation
range that satisfies all the judgment criteria shown in Table 7. X1: κ-carrageenan = 1.0 g, X2: 390
ι-carrageenan = 35 g, X3: oxidized starch = 14 g, and X4: heat-treated starch = 50 g, represents the
central part of the optimal formulation range, and this was determined as the optimal shell
formulation for plant-derived soft gel capsules.
16
395
Fig. 5. The design space meeting the criteria for tensile strength, Young’s modulus, and
adhesive strength.
3.3 Preparation of the optimum formulation of the plant-derived soft gel capsule using the rotary die
method 400
The optimized plant-derived soft gel capsules were prepared using the commercial rotary
die method under manufacturing conditions equivalent to those of existing formulations. The
optimized plant-derived soft gel capsules could be manufactured stably using the rotary die method.
The appearance of pre-existing plant-derived soft gel capsules and the optimal plant-derived soft gel
capsules is shown in Fig. 6. The optimal soft gel capsules had a smooth and transparent appearance 405
comparable to existing ones. As shown in Table 8, the size of the optimal soft gel capsules showed a
relatively high agreement with the specified size of the mold, even after drying.
17
410
Fig. 6. The appearance of the plant-derived soft gel capsules (a: pre-existing, b: optimum
formulation).
The mass of the soft gel capsule content was approximately 200 mg, and this is the ideal
amount for the used mold (Table 8). The mass of the shell was 122 mg for the pre-existing capsule 415
and 93.4 mg for the optimal capsule, suggesting that the optimal soft gel capsules are a more
miniaturized preparation.
The disintegration times for the pre-existing and optimal capsules were 16.1 and 8.5 min,
respectively. The disintegration time of a soft gel capsule is defined as within 20 min in the JP17.
The optimal soft gel capsule showed a very fast disintegration. This might be explained by a 420
decrease in the mass of shell (Table 8) and the thickness of the shell (Fig. 8). Generally, when the
soft gel capsule is disintegrated, encapsulated oily liquids or semi-liquids would be simultaneously
dispersed. Therefore, the optimal capsule with fast disintegration would be useful to exert drug
activity in vivo.
After incisions were made in the soft gel capsule, the capsule wall and the adhesion were 425
observed and this is shown in Fig. 7, while their thickness is shown in Fig. 8. In the optimal soft gel
capsule, the capsule wall and the adhesion were approximately half as thick as those of the
pre-existing capsule. However, as shown in Table 8, the optimal soft gel capsule had the same
hardness as the pre-existing soft gel capsule. From the above results, it is demonstrated that the
tensile strength, flexibility and adhesive strength of the optimal soft gel capsule were increased 430
compared with those of a pre-existing capsule.
18
Fig. 7. The adhesive parts of the plant-derived soft gel capsules (a: pre-existing, b: optimum
formulation). 435
Fig. 8. The thickness of the plant-derived soft gel capsules (a: sheet, b: adhesive part).
Each bar represents the mean ± S.D. (n = 10). **p<0.01 compared with pre-existing.
440
19
Conclusions
In the present study, to optimize a soft gel capsule formulation and reduce the time and cost for the
formulation development of soft gel capsules, a new scale-down methodology, from commercial
rotary die scale to laboratory scale, that can predict the physical properties of soft gel capsules on a 445
commercial scale, was established; i) using the formula information of soft gel capsules produced on
a commercial scale, plant-derived soft gel film sheets were prepared using an applicator on a
laboratory scale and their physicochemical properties were evaluated; ii) using a D-optimal mixture
design, film sheets were prepared by varying their composition, and the physicochemical properties
of the sheets were evaluated. The relationship between composition and variables was determined 450
using a multiple regression analysis and the formulation was optimized; iii) using the optimized
formulation, soft gel capsules were manufactured using a rotary die method to show the feasibility of
our scale-down method. The optimized soft gel capsules showed equivalent or superior physical
properties compared with pre-existing soft gel capsules. In summary, we have developed a new
scale-down methodology to optimize plant-derived soft gel capsules on a commercial scale. Time 455
and cost can be reduced when optimizing and developing soft gel capsules using new raw materials,
because this scale-down method can predict the physical properties of the film sheet at a laboratory
scale.
20
References 460
Badii, F., Howell, N.K., 2006. Fish gelatin: Structure, gelling properties and interaction with egg albumen
proteins. Food Hydrocoll. 20, 630–640. doi:10.1016/j.foodhyd.2005.06.006
Bottom, C.B., Clark, M., Carstensen, J.T., 1997. Dissolution testing of soft shell capsules-acetaminophen
and nifedipine. J. Pharm. Sci. 86, 1057–1061. doi:10.1021/js960263k
Cuppoletti, J., Malinowska, D.H., Tewari, K.P., Li, Q.-J., Sherry, A.M., Patchen, M.L., Ueno, R., 2004. 465
SPI-0211 activates T84 cell chloride transport and recombinant human ClC-2 chloride currents. Am.
J. Physiol. Cell Physiol. 287, C1173–C1183. doi:10.1152/ajpcell.00528.2003
Fagiolino, P., V??zquez, M., Ibarra, M., Magallanes, L., Guevara, N., Fotaki, N., 2014. Sex- and
smoke-related differences in gastrointestinal transit of cyclosporin A microemulsion capsules. Eur.
J. Pharm. Sci. 63, 140–146. doi:10.1016/j.ejps.2014.07.006 470
Gullapalli, R.P., 2010. Soft gelatin capsules (softgels). J. Pharm. Sci. doi:10.1002/jps.22151
Gullapalli, R.P., Mazzitelli, C.L., 2017. Gelatin and Non-Gelatin Capsule Dosage Forms. J. Pharm. Sci.
doi:10.1016/j.xphs.2017.02.006
Homar, M., Suligoj, D., Gasperlin, M., 2007. Preparation of microcapsules with self-microemulsifying
core by a vibrating nozzle method. J. Microencapsul. 24, 72–81. doi:10.1080/02652040601058525 475
Jannin, V., Chevrier, S., Michenaud, M., Dumont, C., Belotti, S., Chavant, Y., Demarne, F., 2015.
Development of self emulsifying lipid formulations of BCS class II drugs with low to medium
lipophilicity. Int. J. Pharm. 495, 385–392. doi:10.1016/j.ijpharm.2015.09.009
Kagamiyama, M., Masuda, Y., Hirayama, T., The Study on Application of Scheffé’s Model to
Optimization of Paint Formulation and Paint Application 2003, Dec No.141 480
Kato, K., Watanabe, K., Inoue, T., Masuda, K., Japanese patent application 2013-040217
Kamiya, S., Nagae, K., Hayashi, K., Suzuki, N., Hayakawa, E., Kato, K., Sonobe, T., Nakashima, K.,
2014. Development of a new evaluation method for gelatin film sheets. Int. J. Pharm. 461, 30–33.
doi:10.1016/j.ijpharm.2013.11.012
Kiefer, J.C., 1992. Optimum Experimental Designs, in: Breakthroughs in Statistics. pp. 400–436. 485
doi:10.1007/978-1-4612-0919-5_28
Lipinski, C., 2002. Poor aqueous solubility - An industry wide problem in drug discovery. Am. Pharm.
Rev. 5, 82–85.
Liu, J., Zhan, X., Wan, J., Wang, Y., Wang, C., 2015. Review for carrageenan-based pharmaceutical
biomaterials: Favourable physical features versus adverse biological effects. Carbohydr. Polym. 490
121, 27–36. doi:10.1016/j.carbpol.2014.11.063
21
Maki, K.C., Ndife, L.I., Kelley, K.M., Lawless, A.L., Brooks, J.R., Wright, S.B., Shields, J.M., Dicklin,
M.R., 2012. Absorption of Folic Acid from a Softgel Capsule Compared to a Standard Tablet. J.
Acad. Nutr. Diet. 112, 1062–1067. doi:10.1016/j.jand.2012.03.001
Misic, Z., Muffler, K., Sydow, G., Kuentz, M., 2012. Novel starch-based PVA thermoplastic capsules for 495
hydrophilic lipid-based formulations. J. Pharm. Sci. 101, 4516–4528. doi:10.1002/jps.23315
Morrison, N.A., Clark, R.C., Chen, Y.L., Talashek, T., Sworn, G., 1999. Gelatin alternatives for the food
industry, in: Physical Chemistry and Industrial Application of Gellan Gum. pp. 127–131.
doi:10.1007/3-540-48349-7_19
Nazzal, S., Wang, Y., 2001. Characterization of soft gelatin capsules by thermal analysis. Int. J. Pharm. 500
230, 35–45. doi:10.1016/S0378-5173(01)00856-0
Piculell, L., Nilsson, S., Muhrbeck, P., 1992. Effects of small amounts of kappa-carrageenan on the
rheology of aqueous iota-carrageenan. Carbohydr. Polym. 18, 199–208.
doi:10.1016/0144-8617(92)90064-W
Reich, G., 2004. Formulation and physical properties of soft capsules. Pharm. Capsul. 201–212. 505
Scheffe, H., 1958. Experiments with Mixtures. J. R. Stat. Soc. Ser. B 20, 344–360.
doi:10.1198/004017002320256620
Scheffé, H., 1963. The Simplex-Centroid Design for Experiments With Mixtures. J. R. Stat. Soc. Ser. B
25, 235–263.
Schrieber, R., Gareis, H., 2007. Introduction. Gelatine Handb. 1–44. doi:10.1002/9783527610969.ch1 510
Shah, N.H., Stiel, D., Infeld, M.H., Railkar A.S., Malick A.W., Patrawala M., 1992. Elasticity of soft
gelatin capsules containing polyethylene glycol 400-quantitation and resolution,Pharm. Technol.,
March, 126–132.
Stanley, J.P., 1986. Soft gelatin capsules. In: Lachman, L., Lieberman, H.A., Kanig, J.L. (Eds.), The
Theory and Practice of Industrial Pharmacy, Third ed. Lea and Febiger, Philadelphia, PA, pp. 398–515
412.
Tesconi, M.S., Bramer, S.L., Yalkowsky, S.H., 1999. The preparation of soft gelatin capsules for a
radioactive tracer study. Pharm. Dev. Technol. 4, 507–513. doi:10.1081/PDT-100101388
Torres, M.D., Chenlo, F., Moreira, R., 2016. Rheology of κ/ι-hybrid carrageenan from Mastocarpus
stellatus: Critical parameters for the gel formation. Int. J. Biol. Macromol. 86, 418–424. 520
doi:10.1016/j.ijbiomac.2016.01.096
22
Table captions
Table 1. Pre-existing formulation of the soft gel capsules. 525
Table 2. Normalized levels of four design variables.
Table 3. Percentage composition (100%) and physical properties.
530
Table 4. Analysis of variance (ANOVA) for tensile strength (Y1).
Table 5. Analysis of variance (ANOVA) for Young’s modulus (Y2).
Table 6. Analysis of variance (ANOVA) for adhesive strength (Y3). 535
Table 7. The criteria of tensile strength, Young’s modulus, and adhesive strength.
Table 8. Physicochemical properties of the optimized soft gel capsules manufactured using the
rotary die method. 540
23
Table 1.
Gelatin Plant-derived
Gelatin (g) 100.0 -
κ-carrageenan (g) - 0.1
ι-carrageenan (g) - 29.9
Oxidized starch (g) - 30.0
Heat-treated starch (g) - 40.0
Glycerin (g) 40.0 50.0
Water (g) 100.0 165.0
Glycerin fatty acid ester (μL) 200 200
Total (g) 240.0 315.0
545
Table 2.
Design variable (g) Low level High level
X1: κ-carrageenan 0 3
X2: ι-carrageenan 20 35
X3: Oxidized starch 0 80
X4: Heat-treated starch 0 80
550
555
24
Table 3.
Batch
No. Factor Response
κ-carrageenan
(g)
ι-carrageenan
(g)
Oxidized
starch
(g)
Heat-treated
starch
(g)
Tensile
strength
(MPa)
Young’s
modulus
(MPa)
Adhesive
strength
(MPa)
X1 X2 X3 X4 Y1 Y2 Y3
1 0 20 80 0 5.35 54.92 5.49
2 0 27.5 72.5 0 4.04 24.12 4.23
3 0 35 65 0 4.56 31.02 4.39
4 0 20 40 40 2.02 10.00 2.97
5 0 35 0 65 3.99 10.03 5.48
6 0 20 0 80 2.35 6.63 3.61
7 1.5 27.5 0 71 3.45 14.00 4.65
8 1.5 20 39.25 39.25 5.15 33.97 4.16
9 1.5 35 63.5 0 5.20 26.49 5.56
10 3 20 77 0 7.49 67.05 5.79
11 3 35 62 0 5.43 30.27 4.66
12 3 27.5 34.75 34.75 5.66 32.79 5.44
13 3 35 0 62 6.05 25.64 5.48
14 3 20 0 77 5.70 36.23 5.46
15 1.5 27.5 35.5 35.5 4.14 23.49 4.74
16 1.5 27.5 35.5 35.5 4.78 25.58 4.86
17 1.5 27.5 35.5 35.5 3.54 22.76 5.06
560
25
Table 4.
Source Sum of squares d.f.a Mean square F-value P-value
Model 27.15 9 3.02 8.06 0.0059
Linear Mixture 19.86 3 6.62 17.69 0.0012
X1X2 0.006 1 0.006 0.016 0.9023
X1X3 0.059 1 0.059 0.16 0.7024
X1X4 0.068 1 0.068 0.18 0.6820
X2X3 1.27 1 1.27 3.40 0.1079
X2X4 0.92 1 0.92 2.45 0.1618
X3X4 0.60 1 0.60 1.59 0.2473
Residual 2.62 7 0.37
Lack of Fit 1.85 5 0.37 0.96 0.5805
Corrected Total 29.77 16
R2 = 0.91, Rf
2 = 0.80, adequate precision = 12.375, corresponding chance ‘Lack of Fit F-value’
= 58.05%
a d.f. indicates degree of freedom.
565
570
575
26
Table 5.
Source Sum of squares d.f.a Mean square F-value P-value
Model 3518.76 9 390.97 11.06 0.0023
Linear Mixture 2677.39 3 892.46 25.24 0.0004
X1X2 2.82 1 2.82 0.080 0.7860
X1X3 10.76 1 10.76 0.30 0.5983
X1X4 12.09 1 12.09 0.34 0.5771
X2X3 94.71 1 94.71 2.68 0.1457
X2X4 60.62 1 60.62 1.71 0.2318
X3X4 107.82 1 107.82 3.05 0.1243
Residual 247.51 7 35.36
Lack of Fit 243.23 5 48.65 22.71 0.0427
Corrected Total 3766.28 16
R2 = 0.93, Rf
2 = 0.85, adequate precision = 14.454, corresponding chance ‘Lack of Fit F-value’
= 4.27%
a d.f. indicates degree of freedom. 580
585
590
595
27
Table 6.
Source Sum of squares d.f.a Mean square F-value P-value
Model 9.25 13 0.71 6.51 0.0744
Linear Mixture 3.44 3 1.15 10.51 0.0423
X1X2 0.013 1 0.013 0.12 0.7551
X1X3 0.33 1 0.33 3.04 0.1797
X1X4 0.36 1 0.36 3.28 0.1678
X2X3 0.95 1 0.95 8.69 0.0602
X2X4 0.71 1 0.71 6.53 0.0835
X3X4 1.90 1 1.90 17.38 0.0251
X1X2X3 0.11 1 0.11 0.97 0.3978
X1X2X4 0.080 1 0.080 0.74 0.4539
X1X3X4 0.094 1 0.094 0.86 0.4217
X2X3X4 0.58 1 0.58 5.30 0.1047
Residual 0.33 3 0.11
Lack of Fit 0.28 1 0.28 10.54 0.0832
Corrected Total 9.58 16
R2 = 0.97, Rf
2 = 0.82, adequate precision = 9.596, corresponding chance ‘Lack of Fit F-value’
= 8.32%
a d.f. indicates degree of freedom.
600
605
28
Table 7.
Tensile strength (MPa) > 4.41
Young’s modulus (MPa) < 17.8
Adhesive strength (MPa) > 4.66
Table 8.
Pre-existing Optimum
Major axis (mm) (n = 10) 11.0 ± 0.14 10.7 ± 0.11
Minor axis (mm) (n = 10) 6.96 ± 0.07 6.75 ± 0.05
Mass of content (mg) (n = 10) 198 ± 1.14 202 ± 0.92**
Mass of shell (mg) (n = 10) 122 ± 1.37 93.4 ± 1.28**
Disintegration (min) (n = 6) 16.1 ± 0.93 8.5 ± 0.0
Hardness (kg) (n = 20) 16.0 ± 3.42 17.8 ± 2.31
**p<0.01 compared with pre-existing. 610
