Journal Pre-proof

Investigation of lyophilized formulation susceptible to the ramp rate of shelf

temperature in the primary drying process

Ryo Ohori, Tomomi Akita, Chikamasa Yamashita

PII: S1773-2247(20)31574-4
DOI: https://doi.org/10.1016/j.jddst.2020.102285
Reference: JDDST 102285

To appear in: Journal of Drug Delivery Science and Technology

Received Date: 28 September 2020

Revised Date: 12 November 2020
Accepted Date: 7 December 2020

Please cite this article as: R. Ohori, T. Akita, C. Yamashita, Investigation of lyophilized formulation
susceptible to the ramp rate of shelf temperature in the primary drying process, Journal of Drug Delivery
Science and Technology, https://doi.org/10.1016/j.jddst.2020.102285.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier B.V.

 1 Investigation of lyophilized formulation susceptible to the ramp rate of shelf temperature in the primary
2 drying process
3
4 Ryo Ohoria,b, Tomomi Akitab, Chikamasa Yamashitab
5
a
6 Pharmaceutical Science and Technology Unit, Eisai Co., Ltd., 1 Kawashimatakehaya-machi, Kakamigahara-
7 shi, Gifu 501-6195, Japan
b
8 Department of Pharmaceutics and Drug Delivery, Faculty of Pharmaceutical Sciences, Tokyo University of
9 Science, 2641 Yamazaki, Noda, Chiba 278-8510, Japan
10

of
11 Corresponding author: Chikamasa Yamashita, Department of Pharmaceutics and Drug Delivery, Faculty of
12 Pharmaceutical Sciences, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278-8510, Japan; Tel:

ro
13 +81-4-7124-1501; Fax: +81-4-7124-3622; e-mail:[email protected]
14

15 ABSTRACT
-p
re
16 In a series of our studies, we revealed that the ramp rate of shelf temperature in the primary drying process
lP

17 affects final product quality in an aggressive cycle (macroscopic collapse occurs at the slow ramp rate whereas
18 acceptable lyophilized cake appearance is observed at the fast ramp rate). However, the studies were performed
na

19 by disaccharide-based formulations, and whether the other formulations are susceptible to the ramp rate
20 remained unclear. In the present study, the aqueous solutions of a representative bulking agent (mannitol,
ur

21 glycine, dextran, sucrose, trehalose and lactose) and protein model formulation with different levels of bovine
22 serum albumin were lyophilized at different ramp rates. Macroscopic collapse occurred at the slow ramp cycle
Jo

23 in a high concentration of disaccharide solutions, which have a low collapse temperature (Tc), and slow ramp-
24 induced collapse was inhibited as the concentration was decreased. Also, drying failure in the slow ramp cycle
25 was not confirmed in mannitol, glycine and dextran formulations. Furthermore, the slow ramp-induced collapse
26 was inhibited as the concentration of the protein was increased. Thus, the amorphous-based formulation with
27 low Tc containing a large amount of bulking agent was considered to be the formulation susceptible to the ramp
28 rate.
29

30 Keywords:

31 Collapse, Collapse temperature, Freeze drying, Lyophilization, Ramp rate

32
33 GRAPHICAL ABSTRACT

of
34

ro
-p
re
lP
na
ur
Jo
35 1. INTRODUCTION

36 In the pharmaceutical industry, lyophilization is an important drying technology used to prepare solid parenteral
37 formulations, and a large number of lyophilized products are commercially provided as injections.
38 Lyophilization is recognized to be a time- and cost-consuming process, and an economical lyophilization
39 program with ensuing product quality is desired. For this reason, great effort is required to establish
40 lyophilization processes, and primary drying conditions, which significantly affect product quality and drying
41 time, are particularly investigated [1–3]. The general strategy to obtain lyophilized product with desired
42 qualities is that the product temperature (Tp) during the primary drying process is maintained below critical
43 temperature to avoid drying failure known as “collapse” or “meltback”. To determine the optimum primary

of
44 drying condition, shelf temperature (Ts) and chamber pressure (Pc), which directly affect heat input into the vial,
45 are mainly investigated, and great number of case studies are available [1–6]. Also, the freezing process affects

ro
46 the drying behavior of the primary drying stage (primary drying time, dry layer resistance and occurrence of
47 drying failure), and freezing conditions have also been focused on [7–11]. Thus, it is generally recognized
48
-p
among scientists and engineers dealing with lyophilization that Ts and Pc for the primary drying process and
re
49 freezing condition are key operational parameters for successful lyophilization.
50 There are some operational parameters other than the Ts and Pc and the ramp rate at the beginning of the
lP

51 primary drying process is one of the minor parameters. Although various ramp rates between 0.03–1.00 °C/min
52 have been applied to prepare lyophilized samples [12–18], there has been no systematic investigation of the
na

53 ramp rate so far, and its effect on product quality is poorly understood. Therefore, we considered it important to
54 clarify the effect of the ramp rate on final product quality to establish successful lyophilization. In our series of
ur

55 previous studies, we revealed that ramp rate is a potentially critical operational parameter. Severe shrinkage and
56 partial melting at the bottom of lyophilized cake were observed in the slow ramp cycle, whereas lyophilized
Jo

57 cake with acceptable cake appearance was obtained in the fast ramp cycle even though same Ts and Pc were
58 applied [19]. Also, we confirmed that the effect of the ramp rate is negligible in a conservative cycle (target Ts =
59 -20 °C) but it becomes observable in the aggressive cycle (target Ts = 0-20 °C). In subsequent studies, the
60 protein formulations prepared by the slow ramp cycle collapsed and showed lower storage stability than that by
61 the fast ramp cycle [20]. Furthermore, investigation of the cause of the collapse induced by the slow ramp cycle
62 showed that the temperature at collapse was different between the fast and slow ramp cycles. We thus concluded
63 that the decrease of collapse temperature (Tc) during the shelf ramp process would be one reason for the
64 collapse [21]. Thus, lyophilization programs may need to take the ramp rate into account.
65 The effect of the ramp rate was observed throughout our studies when the sucrose- and trehalose-based
66 formulations whose Tc is commonly low were lyophilized by aggressive cycle in a range of Ts = 0-20 °C.
67 Recently, several case studies of product prepared by lyophilization cycles with the different ramp rates were
68 reported [22–23]. In these studies, no remarkable effect of the ramp rate could be observed, and an effect
69 opposite to what we claimed (cake appearance was improved in the slow ramp cycle) was confirmed. These
70 studies commonly used high concentrations of protein formulations, which suggests that the occurrence of the
71 change in final product qualities caused by the ramp rate depends on the composition of the formulation in
72 addition to the primary drying condition. Thus, it is possible that the ramp rate is effective in limited
73 formulation. Clarification of the susceptible formulation would be helpful knowledge to avoid slow ramp-
74 induced drying failure and valuable information for establishing a successful lyophilization program. However,
75 our studies were performed using formulations of amorphous disaccharide so far, and whether other
76 formulations with crystalline and polymer bulking agent susceptible to the ramp rate remained unclear.
77 The aim of the present study was to identify the formulation susceptible to the ramp rate of Ts during the
78 primary drying process. In the lyophilized formulation, the bulking agent is the major component and
79 lyophilization properties (drying behavior, physicochemical properties of lyophilized cake, etc.) are often

of
80 predominantly determined by the bulking agent. Therefore, simple solutions of various bulking agents were
81 used to estimate the susceptible formulation to the ramp rate. Representative bulking agents commonly used in

ro
82 lyophilized product for injections (mannitol, glycine, dextran, sucrose, trehalose and lactose) were selected, and
83
84
-p
the solutions containing different levels of bulking agent were lyophilized by slow and fast ramp cycles. Also,
the lyophilized products containing different levels of protein (bovine serum albumin [BSA]) were investigated
re
85 to evaluate whether the degree of effect of the ramp rate is varied by protein quantity. Based on the results
86 obtained, the formulation susceptible to the ramp rate was discussed.
lP
na
ur
Jo
 87 2. MATERIALS AND METHODS

88 2.1 MATERIALS

89 All chemicals used in this study were obtained from the following commercial vendors: BSA (Sigma Aldrich, St.
90 Louis, MO); D-(+)-trehalose dihydrate (Hayashibara, Okayama, Japan); glycine, sucrose, L-arginine, L-arginine
91 hydrochloride and citric acid monohydrate (EMD Millipore, Billerica, MA); Tween 80 (Croda, Edison, NJ);
92 mannitol (Roquette, Lestrem, France); lactose (Avantor, Center Valley, PA); dextran (Meito Sangyo, Aichi,
93 Japan); and Hydranal® Coulomat AG and Hydranal® Coulomat CG (Sigma Aldrich). The packaging materials
94 used in the present study were obtained from the following commercial vendors: 5 mL type 1 glass vials (BB
95 vial 23×43 VIST; outside diameter, 23 mm; inside diameter, 20 mm; Daiwa Special Glass, Osaka, Japan) and

of
96 rubber stoppers (20-mm gray butyl rubber stoppers, V10-F8, D713, RB2-40; Daikyo Seiko, Tokyo, Japan).

ro
97 2.2 SAMPLE PREPARATION AND LYOPHILIZATION CONDITIONS

98
99
-p
Mannitol, glycine, dextran, sucrose, trehalose and lactose were selected as the representative bulking agents for
lyophilized product in the present study (Table 1). Simple solutions containing three different levels (5wt%,
re
100 10wt% and 15wt%) of bulking agents were prepared and lyophilized. BSA was used as model protein because it
101 is widely used in various research fields. Proteins are frequently formulated with non-reducing disaccharide [24-
lP

102 25], and trehalose-based model formulation for protein was applied in our previous study [19]. Therefore,
103 0.1wt%, 0.5wt%, 1.0wt% and 2.5wt% formulated BSA solutions containing mainly 10wt% trehalose were
na

104 prepared and lyophilized. A pilot scale lyophilizer (S20NS, Nissan Edwards, Osaka, Japan) was used for
105 lyophilization of the samples. The sample solution (2.2 mL) was filled in a glass vial, which was then partially
ur

106 stoppered with a rubber stopper. The number of vials filled with the simple solutions of bulking agent and
107 protein formulation were 20 vials and 50 vials, respectively. The number of formulation was large and it was
Jo

108 impractical to obtain reproducible Tp for all formulations due to limited number of thermocouple. For this
109 reason, thermocouple was not set in the filled vial and the samples were lyophilized together in several runs. To
110 avoid the effect of radiative heat (edge vial effect), the product vials were placed at the center of the tray, and
111 vials filled with purified water were placed around the products.
112 With regard to the primary drying condition, the effect of the ramp rate was investigated by the various primary
113 drying conditions and it could be confirmed at less than 0.33 °C/min in the first study when aggressive primary
114 drying (target Ts = 0-20 °C and Pc = 10 Pa) was conducted [19]. For this reason, the effect of the ramp rate was
115 investigated at 1.00 °C/min (fast ramp cycle) and 0.06 °C/min (Slow ramp cycle) with target Ts = 0 °C and Pc =
116 10 Pa as representative drying conditions in the second and third studies [20-21]. It was expected that the
117 obtained results of the present study can be discussed comparing the previous studies when the same primary
118 drying condition is applied. Therefore, the samples were lyophilized by following programs which is same
119 primary drying condition as the second and third studies. (1) Ts was initially attained by cooling from ambient
120 temperature to -40 °C (0.67 °C/min). (2) Ts was then held at -40 °C for 3 h, and (3) Pc was reduced to 10 Pa. (4)
121 Ts was then raised by heating from -40 °C to 0 °C at a ramp rate of 1.00 °C/min (fast ramp cycle) or
122 0.06 °C/min (slow ramp cycle). (5) Ts and Pc were maintained for 24 h. (6) For secondary drying, Ts was
123 changed to 25 °C for 6 h, and (7) Ts and Pc were then maintained for 24 h.

124 2.3 CAKE APPEARANCE

125 Cake appearance was evaluated by visual observation. Severe shrinkage and partial melting at the bottom of the
126 lyophilized cake due to drying failure were assessed. These products were classified as unacceptable cakes
127 (defect), and the defect rate of cake appearance was calculated as a percentage of unacceptable cakes. The cross-
128 sectional appearance of the lyophilized cake was also observed by the method described previously [21]. The
129 cake was carefully removed from the vial by breaking the bottle neck and split to evaluate the inner structure.

of
130 The cross-sectional appearance of the cake was observed using a digital microscope (VHX-5000, Keyence,
131 Osaka, Japan).

ro
132 2.4 SCANNING ELECTRON MICROSCOPY (SEM)

133
134
-p
Evaluation was performed by the method described previously [21]. SEM observation was directly conducted
without metal coating because the SEM (VE-7800, Keyence) used does not require metallization prior to
re
135 observation. The accelerating voltage was 1 kV at a working distance of approximately 4-11 mm.
lP

136 2.5 RESIDUAL MOISTURE

137 Evaluation was performed by the method described previously [20]. The amount of residual moisture was
na

138 evaluated by coulometric Karl Fischer titration (CA-200, Mitsubishi Chemical Analytech, Kanagawa, Japan).
139 Under dry atmosphere, the lyophilized cake was crushed with a spatula. Approximately 50–150 mg of crushed
ur

140 powder was dissolved in anolyte (Hydranal® Coulomat AG) and titrated until the endpoint. Residual moisture
141 (%) was calculated from the water amount determined by the titration and mass of the lyophilized powder.
Jo

142 2.6 RECONSTITUTION TIME

143 Evaluation was performed by the method described previously [20]. The samples of each formulation were
144 reconstituted with 2 mL of WFI (water for injection) and gently agitated for 5 s. The reconstituted vials were
145 visually observed against a black background and compared with the control sample, a vial of WFI. The
146 observation was performed under ambient temperature in an approximately 1000 lux environment and
147 continued until the appearance of the reconstituted solution matched that of WFI. After the gentle agitation
148 applied at reconstitution, no further agitation or shaking were applied during the observation. The time required
149 for the complete dissolution was recorded as the reconstitution time, and measurements were performed in
150 triplicate.

151 2.7 X-RAY POWDER DIFFRACTION (XRD)

152 XRD measurement was performed to confirm crystallization of the products. The measurement was performed
153 by the method described previously [20]. Under dry atmosphere, the lyophilized cake was crushed with a
154 spatula, and the lyophilized powder was put into a sample cell. The samples were measured by an Empyrean X-
155 ray diffractometer (PANalytical, Almelo, Netherlands) with Cu Kα radiation at 40 kV/40 mA. The samples were
156 scanned over a range of 2θ = 5–40° with resolution of 0.01313° and a step time of 24 s.
157

158 3. RESULTS

159 3.1 APPEARANCE AND MICROSTRUCTURE OF LYOPHILIZED CAKES

160 The lyophilized cakes containing different levels of bulking agents were lyophilized by the fast and slow ramp
161 cycles, with representative cake appearances shown in Fig 1. Regardless of the ramp rates, elegant cake

of
162 appearance was confirmed in the Man 100, Gly 100 and Dex 100 formulations (Fig. 1). For the Suc 100, Tre
163 100 and Lac 100 formulations, although the lyophilized products prepared by the fast ramp cycle showed

ro
164 acceptable cake appearance, severe shrinkage was observed at the bottom of the cake in the slow ramp cycle.
165
166
-p
The obtained lyophilized products were visually inspected, and the defect rate of cake appearance are
summarized in Fig. 2. Regardless of the ramp rates and concentration of the bulking agents, no defect of cake
re
167 appearance was observed, and all lyophilized product formed elegant lyophilized cakes in the Man, Gly and
168 Dex formulations (Fig. 2A–C). In contrast, although some lyophilized cakes were shrunk in the fast ramp cycle,
lP

169 the defect rate in the slow ramp cycle was evidently higher than that in the fast ramp cycle for the Suc, Tre and
170 Lac formulations (Fig. 2D–F). Also, the defect rate was consistently increased as the concentrations Suc, Tre
na

171 and Lac were increased in both cycles. All lyophilized cakes of Tre and Lac formulations prepared by the slow
172 ramp cycle showed severe shrinkage or partial melting at a concentration of more than 10wt%, whereas some
ur

173 lyophilized cakes of these formulations formed acceptable cake appearance at 5wt%. All lyophilized cakes of
174 Suc formulations prepared by slow ramp cycle were judged as defective regardless of the concentration.
Jo

175 The cross-sectional appearance of the Suc, Tre and Lac formulations deteriorated with the increase in the
176 concentration, and in the Suc formulation, the inner cake structure of the sample prepared by the slow ramp
177 cycle at the highest concentration (15wt%) was severely melted (Fig. 3). Because there was no difference in the
178 macroscopic appearance of the Man, Gly and Dex formulations, the microstructure of these formulations was
179 observed by SEM. A uniform porous structure was observed, and no remarkable difference in the microstructure
180 between both cycles was confirmed (Fig. 4A–D). In the Suc formulation, although void space due to collapse
181 was observed in the slow ramp cycle, a fine porous structure was confirmed elsewhere. In contrast, larger pore
182 overall was observed in the fast versus the slow ramp cycle formulations. Thus, no difference in cake
183 appearance was observed in the Man, Gly and Dex formulations between the fast and slow ramp cycles,
184 whereas a clear difference was confirmed in the Suc, Tre and Lac formulations.

185 3.2 PHYSICOCHEMICAL PROPERTIES OF LYOPHILIZED CAKES

186 Among the physicochemical properties of the obtained lyophilized cakes evaluated, there was no remarkable
187 difference of the residual moisture between the fast and slow ramp cycles in the Man, Gly and Dex formulations
188 (Fig. 5A–C). In contrast, although the Suc 50, Tre 50 and Lac 50 formulations prepared by both lyophilization
189 cycles showed comparable residual moisture, that in the slow ramp cycle was greater as the concentration
190 increased (Fig. 5D–F). In particular, the residual moisture of the Suc formulation was extremely increased by
191 the slow ramp cycle, and that of the Suc 150 formulation prepared by slow ramp cycle was approximately 3%
192 higher than that of the fast ramp cycle. Likewise, there was no remarkable difference in reconstitution times of
193 the Man, Gly and Dex formulations, which ranged between 30–70 s regardless of the ramp rate (Fig. 6A–C).
194 The reconstitution times of lowest concentrations of the Suc 50, Tre 50 and Lac 50 formulations ranged between
195 20–30 s regardless of the ramp rate (Fig. 6D–F). However, the reconstitution times of the cakes prepared by the
196 slow ramp cycle became longer than those by the fast ramp cycle as the concentration increased.

of
197 XRD was measured to evaluate the crystal structure of the lyophilized cakes. There was no remarkable
198 difference in the diffraction patterns between the two ramp cycles (Fig. 7). Based on knowledge from previous

ro
199 studies [26-27], the crystal forms of the Man 100 and Gly 100 formulations were identified as an α/δ mixture
200
201
-p
and β form, respectively. The diffraction pattern for the Suc 100 formulation showed no crystallin peak (halo
pattern) in either cycle. The Tre 100, Lac 100 and Dex 100 formulations showed similar results. Thus, no effect
re
202 of the ramp rate was observed in the Man, Gly and Dex formulations in terms of physicochemical properties.
203 Although the effect of the ramp rate on the cake appearance was confirmed at all concentration in the Suc, Tre
lP

204 and Lac formulations, a clear difference in the physicochemical properties could not be confirmed at the lowest
205 concentration.
na

206 3.3 EFFECT OF RAMP RATE ON PROTEIN FORMULATIONS

ur

207 To evaluate the effect of the ramp rate on protein formulations, the lyophilized product with different quantities
208 of BSA was prepared by the fast and slow ramp cycles. Although the macroscopic appearance of lyophilized
Jo

209 cakes of the Pro 5 formulation prepared by the slow ramp cycle showed severe shrinkage, an acceptable cake
210 appearance was observed in the Pro 25 formulation (Fig. 8A, D). The cross-sectional appearance of Pro 5 and
211 Pro 25 formulations prepared by the fast ramp cycle was homogenous inner structure (Fig. 8B, C). In contrast,
212 that prepared by the slow ramp cycle indicated void space at the middle layer (Fig. 8E, F). However, the cross-
213 sectional appearance was improved by the increased BSA concentration in the Pro 25 formulation.
214 Although the defect rate of the slow ramp cycle at the lowest BSA concentration was approximately 95%,
215 it gradually decreased as the concentration of BSA increased (Fig. 9A). The defect rate of lyophilized product
216 prepared by the slow ramp cycle at the highest BSA concentration (2.5wt%) was approximately 6%, and the
217 appearance of the cake was remarkably improved. Similarly, the residual moisture and reconstitution time in the
218 lyophilized product prepared by the slow ramp cycle was improved as the concentration of BSA increased (Fig.
219 9B, C). In particular, there was no significant difference in residual moisture and reconstitution times between
220 the fast and slow ramp cycles in the Pro 25 formulations. Thus, the effect of the ramp rate was confirmed to be
221 negligible at a high protein concentration.
222 4. DISCUSSION

223 4.1 EFFECT OF RAMP RATE IN THE LYOPHILIZED PRODUCT

224 The macroscopic appearance of lyophilized cakes of the Suc, Tre, Lac and Pro formulations indicated severe
225 shrinkage and partial melting when the slow ramp cycle was applied. These findings (slow ramp-induced
226 collapses) corresponded with those of our previous studies [21]. It is generally known that collapse or meltback
227 occurs when Tp during primary drying reaches the critical temperature (Tc or Teu). Although various factors (e.g.,
228 Ts, Pc, Rp and heat transfer coefficient) relate to Tp, it is predominantly determined by Ts and Pc. In the present
229 study, Tp was not recorded because many different formulations were lyophilized in several lyophilization runs
230 and reproducible Tp data was unavailable due to the limited number of thermocouples. However, Tp of the

of
231 lyophilized formulation (same formulation as the “Pro 0” and “Pro 1” formulation in the present study) in the
232 same primary drying condition was monitored in our previous studies [18-20] and these Tp data are available to

ro
233 discuss this phenomena. In our series of studies, collapse occurred by slow ramp rate regardless of the fact that
234
235
-p
there was no significant difference of maximum Tp during primary drying process between the fast and slow
ramp cycles. Two hypotheses may explain the difference between the ramp rate: (1) generation of
re
236 “microcollapse” in the fast ramp cycle, and (2) change in the Tc during the shelf ramp process.
237 The microstructure of the dried product (product resistance: Rp) is a one of the factors that affects Tp, and
lP

238 the generation of microcollapse decreases Rp [28-30]. In the SEM image of lyophilized solid for 10% Suc,
239 slightly larger pores were observed in the fast ramp cycle compared to those in the slow ramp cycle (porous
na

240 structure other than at the point of collapse). The rapid Tp increase during the fast ramp cycle caused
241 microcollapse; as a result, the improved sublimation rate reduced Tp which is kept below Tc. In contrast, the
ur

242 slow Tp increase during the slow ramp cycle did not cause microcollapse; as a result, Rp kept increase
243 continuously throughout primary drying until Tp > Tc. For this reason, the macroscopic collapse may be induced
Jo

244 in the slow ramp cycle.

245 With regard to the second hypothesis, Tc is a dynamic parameter and depend on the drying situation. This
246 hypothesis was derived from a previous study from Pikal and Shah [31]. In their study, the sublimation rate,
247 which relates to the desorption rate of water from the dried region, was one of the important parameters
248 affecting theoretical Tc (Tc increases as the sublimation rate is increased). Although there are other several
249 potential factors affecting Tc, their study suggests that Tc during the real primary drying process may be slightly
250 changed by the sublimation rate. A low sublimation rate is maintained for a long time in the slow ramp cycle
251 because the sublimation rate is gradually increased until Ts reaches the target value (detailed explanation for the
252 hypothesis is described in our published paper [20]). In contrast, a high sublimation rate is immediately
253 maintained in the fast ramp cycles because Ts rapidly reaches the target value and microcollapse is induced.
254 Accordingly, Tc during shelf ramp process is changed by the ramp rate, and risk of macroscopic collapse in the
255 fast ramp cycle is expected to be lower than that in the slow ramp cycle. In our previous experimental LT-FDM
256 data, Tc changed by approximately 3 °C at a ramp rate between 0.10 °C/min and 1.00 °C/min [21].
257 These hypotheses indicate that the effect of the ramp rate becomes observable in the aggressive primary
258 drying condition in which Tp is close to the variable range of Tc. Fig. 10 schematically illustrates the
259 relationship between Tc and Tp. When the conservative cycle is selected (case 3), Tp during the shelf ramp
260 process is maintained sufficiently lower than the variable range of Tc. Macroscopic collapse and microcollapse
261 does not occur regardless of the ramp rate, and no effect of the ramp rate is observed. Similarly, no effect of the
262 ramp rate is observed in the further aggressive cycle (case 1) because Tp reaches a higher value than the variable
263 range of Tc in both of the cycles, and macroscopic collapse occur regardless of the ramp rate. Thus, the effect of
264 the ramp rate become observable when the specific primary drying condition is applied (case 2). This effect is
265 further depending on the formulation. There was no remarkable change in macroscopic appearance and
266 physicochemical properties of the lyophilized products between the fast and slow ramp cycles in the mannitol,

of
267 glycine and dextran formulations. A high critical temperature is thought to be one of the reasons why drying
268 failure did not occur in the slow ramp cycle. The critical temperature during the primary drying process of these

ro
269 excipients is commonly high (Tc for Dextran is -10 °C, and the Teu [eutectic melting temperature] for Mannitol
270
271
-p
and Glycine are -1.4 °C and -3.6 °C, respectively [26-27,32] (Table 2). In the present study, the Tp was not
recorded because of the abovementioned reasons, Tp during the primary drying process in previous studies was
re
272 between approximately -25 °C and -20 °C [22,33], even when the target Ts for primary drying was set to 30 °C.
273 Therefore, it is impossible for Tp during the primary drying to approach the critical temperature of these
lP

274 excipients within the primary drying condition conventionally applied. For the dextran formulations, Tp would
275 be sufficiently lower than Tc even when the decrease of Tc by the slow ramp cycle is considered. The same
na

276 consideration can be applied to Man and Gly formulations. Thus, the effect of the ramp rate could not be
277 observed in these formulations (Fig. 10, case 3). We have not investigated the effect of the ramp rate in the
ur

278 lyophilized crystalline solid so far. Drying failure (Meltback) of lyophilized crystalline solid will occur without
279 viscous flow of amorphous solid and its mechanism is different from that of an amorphous solid. Therefore, the
Jo

280 above two hypotheses would not be applicable to crystalline solids (mannitol and glycine formulations). It may
281 be pointed out that the effect of the ramp rate should be investigated by more aggressive primary drying
282 condition. However, it would be concluded that the effect of the ramp rate is “practically” negligible.
283 In the present study, it was revealed that the slow ramp-induced collapse in the Suc, Tre and Lac
284 formulations become remarkable as the concentration of bulking agent is increased. Experimental data from
285 Mesiter et al. indicated that the Tc measured by LT-FDM increased with an increase in the concentration of
286 excipient (2-(hydroxypropyl)-beta-cyclodextrin (HP-β-CD) and polyvinylpyrrolidone (PVP), in a range at 5-
287 15wt%) [34]. Similar results were obtained in lactose solution [35]. Thus, we initially predicted that the slow
288 ramp-induced collapse is avoided at high concentrations, because the decrease of the Tc during shelf ramp
289 process in the slow ramp cycle is offset by the increase in the Tc. However, the study results showed a different
290 trend. Meister at al mentioned in the literature that there is an observation problem in the high concentration
291 solution [34], and it is expected that Tc is not changed by the concentration. The dry-layer resistance (Rp) is
292 increased as the solute concentration is increased in the previous study [36], and Tp is theoretically affected by
293 Rp [4,37]. Although no Tp data is available in the present study, the increase of the Tp is considered to be the
294 most reasonable cause for the severe collapse in the high-concentration region.
295 In the protein formulations containing BSA, product collapse induced by the slow ramp cycle was
296 improved as the concentration of BSA was increased. Particularly, the protein formulation containing highest
297 BSA quantity whose weight ratio of BSA to trehalose was 0.25 (BSA/trehalose) remarkably inhibited the slow
298 ramp-induced collapse. It is generally known that Tc measured by LT-FDM increases as protein quantity is
299 increased. According to the results from Meister and Gieseler [38], Tc for a binary mixture of BSA and trehalose
300 at a BSA/trehalose weight ratio of 0.25 was several degrees higher than that of pure trehalose. In our previous
301 study, there was approximately 3-4 °C of difference for Tc between 0.1% and 2.5% BSA formulation (these
302 were formulated with 10% sucrose and 0.01% polysorbate 80 in pH 5.5, 10 mM citrate buffer) [19]. Also,

of
303 several studies showed that product collapse was avoided by the increase in protein concentration. Depaz et al.
304 [39] reported that formulations of IgG1 and IgG4 were successfully lyophilized at a high protein concentration

ro
305 by an aggressive cycle, whereas collapse was observed at a low protein concentration. Furthermore, other
306
307
-p
studies suggested that collapse does not occur in a high-concentration mAb formulation even when Tp during
the primary drying process reaches its Tc [1,33]. The addition of a large amount of protein increased Tc more
re
308 than the decrease in the Tc due to the slow ramp did. Accordingly, the situation in case 2 of Fig. 10 changed to
309 case 3 as the BSA concentration was increased. Thus, although the effect of the ramp rate might be observable
lP

310 in a more aggressive cycle (high target Ts with fast and slow ramp rates), that in the high-concentration protein
311 formulation would be practically negligible in a general lyophilization program.
na

312 4.2 SUSCEPTIBLE FORMULATION TO THE RAMP RATE

ur

313 To evaluate the effect of the ramp rate comprehensively, we focused on the bulking agent, which is a main
314 component in lyophilized product and prepared the lyophilized products with various bulking agents. Among
Jo

315 the commercial lyophilized products for small molecules, mannitol, lactose and glycine are popular bulking
316 agents [40]. Also, polymer excipients such as dextran are often added to formulations to increase critical
317 temperature (Tc) during primary drying [32]. Sucrose and trehalose are frequently added to lyophilized protein
318 for stabilization [24-25, 41]. The bulking agent for lyophilized product is generally added in a range of 1–
319 10wt%. In particular, the quantity of non-reducing disaccharides used for commercial protein formulations is in
320 a range of 4–160 mg/mL [41]. Regarding protein formulation, commercial lyophilized products with high
321 protein concentration >100 mg/mL are rare, and a protein concentration <50 mg/mL is popular [41]. For these
322 reasons, the examined bulking agents and the concentrations for the component applied in the present study
323 were expected to be reasonable.
324 Based on the obtained results and the abovementioned discussion, the following points can be summarized:
325 (1) the effect of the ramp rate is practically negligible in mannitol and glycine formulations, (2) amorphous-
326 based lyophilized solids with low Tc such as sucrose, trehalose and lactose formulations are affected by the
327 ramp rate in an aggressive cycle, (3) the effect of the ramp rate becomes remarkable as the concentration of
328 bulking agent is increased, and (4) the effect of the ramp rate becomes negligible as the concentration of protein
329 is increased. Thus, an amorphous-based formulation with low Tc is considered to be a formulation susceptible to
330 the ramp rate. Two practical, probable formulations susceptible to the ramp rate can be given as examples: (a)
331 lyophilized protein formulated with more than 10wt% of non-reducing disaccharides at low protein
332 concentration, and (b) small quantity of low molecular drug formulated with more than 10wt% of disaccharides.
333 Among lyophilized formulations, product is often formulated with a mixture of bulking agents. In
334 particular, protein is often formulated with mixture of mannitol and sucrose [24] because an elegant lyophilized
335 cake is obtained by formation of the crystalline scaffold of mannitol. However, the protein in a crystalline solid
336 is unstable, and the addition of large amount of mannitol to a sucrose-based protein formulation results in
337 destabilization [42]. Another previous study reported a desired product that ensures the storage stability of mAb

of
338 and cake elegance is obtained when the mannitol to sucrose ratio is 4:1 [43]. In that case, the boundary
339 temperature that induces drying failure is shifted to the Teu for mannitol (i.e., -1.4 °C) as the ratio of mannitol

ro
340 become larger. Therefore, the effect of the ramp rate would be practically negligible because of the reason
341
342
discussed in section 4.2. -p
When a polymer excipient is added to an amorphous lyophilized solid, the effect of the ramp rate is
re
343 expected to be negligible as the quantity of polymer is increased. Haeuser et al. [22] prepared lyophilized mAb
344 (10 mg/mL and 50 mg/mL) formulated with 2-(hydroxypropyl)-beta-cyclodextrin (HP-β-CD) and Suc that was
lP

345 lyophilized by an aggressive cycle (Ts = 30 °C, Pc = 100 mTorr) with fast and slow ramp cycles (1.0 and
346 0.2 °C/min). There was no significant different in appearance, residual moisture and protein stability between
na

347 the fast and slow ramp cycles [22]. It is expected that Tp in the study of Haeuser et al. was maintained
348 sufficiently lower than its Tc (Fig. 10, case 3) because Tc was increased by the addition of HP-β-CD. Therefore,
ur

349 the effect of the ramp rate could not be observed.

Jo

350 4.3 COMPARISON TO PREVIOUS STUDIES

351 In a recent study of lyophilization for protein formulations, a single-step approach (secondary drying is omitted
352 by performing aggressive primary drying) has emerged [22-23]. This approach was successful due to the
353 increase of Tc by addition of high concentrations of protein and polymer bulking agent. In these studies, the
354 effect of ramp rate on final product quality was investigated. Haeuser et al. [22] could not confirm the effect of
355 the ramp rate, and its cause is discussed in above in section 4.3. In the study of Pansare and Patel [23], the
356 shrinkage of lyophilized cake, which consisted of 50 mg/mL of mAb, 7-8wt% disaccharide and surfactant, was
357 improved by a decrease of the slow ramp rate. This result indicates that a slow ramp rate can positively affect
358 cake appearance, which is the opposite phenomena that we claim. In their study, the target Ts was much higher
359 (the effect of the ramp rate was investigated at Ts = 60 °C, ramp rate: 0.5 and 0.1 °C/min) than the Ts in the
360 present study, and this target Ts is not usually applied in conventional lyophilization programs. Although more
361 detailed data regarding the ramp rate was not shown in that study, it is expected that the duration of Tp
362 maintained below Tc in the slow ramp cycle was longer than that in the fast ramp cycle. Thus, a slow ramp rate
363 may be better than a fast ramp cycle to ensure final product quality in a cycle using an extremely high Ts (Fig.
364 10, case 1).

of
ro
-p
re
lP
na
ur
Jo
365 5. CONCLUSION

366 To identify a formulation susceptible to the ramp rate, a large number of formulations containing different levels
367 of bulking agent was lyophilized by fast and slow ramp cycles. As a result, the following points were clarified:
368 (1) the effect of the ramp rate is practically negligible in mannitol and glycine formulations, (2) an amorphous-
369 based lyophilized solid with low Tc is affected by the ramp rate, (3) the effect of the ramp rate becomes
370 remarkable as the concentration of bulking agent is increased, and (4) the effect of the ramp rate becomes
371 negligible as the concentration of protein is increased. In summary, an amorphous-based formulation with low
372 Tc containing a large amount of bulking agent was considered to be the formulation most susceptible to the
373 ramp rate. Accordingly, the ramp rate would affect the final product quality in a specific formulation. When a
374 formulation similar to the above formulation is lyophilized, the primary drying condition should be designed

of
375 taking into account the effect of the ramp rate.

ro
376

377 6. CREDIT AUTHORSHIP CONTRIBUTION STATEMENT

378
-p
Ryo Ohori: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing - original draft,
re
379 Writing - review & editing, Visualization. Tomomi Akita: Writing - review & editing. Chikamasa Yamashita:
380 Writing - review & editing, Supervision, Project administration.
lP

381

382 7. DECLARATION OF INTEREST

na

383 The authors declare that they have no known competing financial interests or personal relationships that could
ur

384 have appeared to influence the work reported in this paper.

Jo
385 8. REFERENCES

386 1. J.D. Colandene, L.M. Maldonado, A.T. Creagh, J.S. Vrettos, K.G. Goad, T.M. Spitznagel, Lyophilization
387 cycle development for a high-concentration monoclonal antibody formulation lacking a crystalline
388 bulking agent. J. Pharm. Sci. 96 (2007) 1598–1608.
389 2. T. Kramer, D.M. Kremer, M.J. Pikal, W.J. Petre, E.Y. Shalaev, L.A. Gatlin, A procedure to optimize
390 scale-up for the primary drying phase of lyophilization. J. Pharm. Sci. 98 (2009) 307–318.
391 3. S.C. Tsinontides, P. Rajniak, D. Pham, W.A. Hunke, J. Placek, S.D. Reynolds, Freeze drying--Principles
392 and practice for successful scale-up to manufacturing. Int. J. Pharm. 280 (2004) 1–16.
393 4. T. Kodama, H. Sawada, H. Hosomi, M. Takeuchi, N. Wakiyama, E. Yonemochi, K. Terada, Optimization

of
394 of primary drying condition for pharmaceutical lyophilization using a novel simulation program with a
395 predictive model for dry layer resistance. Chem. Pharm. Bull. 62 (2014) 153–159.

ro
396 5. V.R. Koganti, E.Y. Shalaev, M.R. Berry, T. Osterberg, M. Youssef, D.N. Hiebert, F. Kanka, M. Nolan, R.
397 Barrett, G. Scalzo, G. Fitzpatrick, N. Fitzgibbon, S. Luthra, L. Zhang, Investigation of design space for
398
-p
freeze-drying: use of modeling for primary drying segment of a freeze-drying cycle. AAPS
re
399 PharmSciTech. 12 (2011) 854–861.
400 6. S.M. Patel, M.J. Pikal, Lyophilization process design space. J. Pharm. Sci. 102 (2013) 3883–3887.
lP

401 7. G. Assegehegn, E.B. Fuente, J.M. Franco, C. Gallegos. The importance of understanding the freezing step
402 and its impact on freeze drying process performance. J. Pharm. Sci. 108 (2019) 1378–1395.
na

403 8. R. Esfandiary, S.K. Gattu, J.M. Stewart, S.M. Patel. Effect of Freezing on Lyophilization Process
404 Performance and Drug Product Cake Appearance. J. Pharm. Sci. 105 (2016) 1427–1433.
ur

405 9. R. Geidobler, G. Winter, Controlled ice nucleation in the field of freeze-drying: Fundamentals and
406 technology review. Eur. J. Pharm. Biopharm. 85 (2013) 214–222.
Jo

407 10. J.C. Kasper, W. Friess, The freezing step in lyophilization: Physico-chemical fundamentals, freezing
408 methods and consequences on process performance and quality attributes of biopharmaceuticals. Eur. J.
409 Pharm. Biopharm. 78 (2011) 248–263.
410 11. J.A. Searles, J.F. Carpenter, T.W. Randolph, Annealing to optimize the primary drying rate, reduce
411 freezing-induced drying rate heterogeneity, and determine T’g pharmaceutical lyophilization. J. Pharm.
412 Sci. 90 (2001) 872–887.
413 12. A. Clausi, P. Chouvenc, Formulation approach for the development of a stable, lyophilized formaldehyde-
414 containing vaccine. Eur. J. Pharm. Biopharm. 85 (2013) 272–278.
415 13. T. Guan, Y. Miao, L. Xu, S. Yang, J. Wang, H. He, X. Tang, C. Cai, H. Xu, Injectable nimodipine-loaded
416 nanoliposomes: Preparation, lyophilization and characteristics. Int. J. Pharm. 410 (2011) 180–187.
417 14. V.P. Heljo, H. Harju, T. Hatanpää, G. Yohannes, A.M. Juppo, The effect of freeze-drying parameters and
418 formulation composition on IgG stability during drying. Eur. J. Pharm. Biopharm. 85 (2013) 752–755.
419 15. T.H. Jin, L. Nguyen, T. Qu, E. Tsao, Improved formulation and lyophilization cycle for rBCG vaccine.
420 Vaccine. 29 (2011) 4848–4852.
421 16. S. Kadoya, K. Fujii, K. Izutsu, E. Yonemochi, K. Terada, C. Yomota, T. Kawanishi, Freeze-drying of
422 proteins with glass-forming oligosaccharide-derived sugar alcohols. Int. J. Pharm. 389 (2010) 107–113.
423 17. J.D. Meyer, R. Nayar, M.C. Manning, Impact of bulking agents on the stability of a lyophilized
424 monoclonal antibody. Eur. J. Pharm. Sci. 38 (2009) 29–38.
425 18. J.C. Zillies, K. Zwiorek, F. Hoffmann, A. Vollmar, T.J. Anchordoquy, G. Winter, Formulation
426 development of freeze-dried oligonucleotide-loaded gelatin nanoparticles. Eur. J. Pharm. Biopharm. 70
427 (2013) 514–521.
428 19. R. Ohori, C. Yamashita, Effects of temperature ramp rate during the primary drying process on the
429 properties of amorphous-based lyophilized cake, Part 1: Cake characterization, collapse temperature and

of
430 drying behavior. J. Drug Deliv. Sci. Technol. 39 (2017) 131–139.
431 20. R. Ohori, T. Akita, C. Yamashita, Effect of temperature ramp rate during the primary drying process on

ro
432 the properties of amorphous-based lyophilized cake, Part 2: successful lyophilization by adopting a fast
433
434
-p
ramp rate during primary drying in protein formulations. Eur. J. Pharm. Biopharm. 130 (2018) 83–95.
21. R. Ohori, T. Akita, C. Yamashita, Mechanism of collapse of amorphous-based lyophilized cake induced
re
435 by slow ramp during the shelf ramp process. Int. J. Pharm. 564 (2019) 461–471.
436 22. C. Haeuser, P. Goldbach, J. Huwyler, W. Friess, A. Allmendinger, Be aggressive! Amorphous excipients
lP

437 enabling single-step freeze drying of monoclonal antibody formulations. Pharmaceutics 11 (2019) 616.
438 23. S.K. Pansare, S.M. Patel, Lyophilization process design and development: A single-step drying approach.
na

439 J. Pharm. Sci. 108 (2019) 1423–1433.

440 24. B.K. Meyer, M. Shameem, Commercial therapeutic protein drug products, Therapeutic Protein Drug
ur

441 Products. Woodhead Publishing Limited. (2012) 1–11.

442 25. W. Wang, Lyophilization and Development of Solid Protein Pharmaceuticals. Int. J. Pharm. 203 (2003)
Jo

443 1–60.
444 26. A.I. Kim, M.J. Akers, S.L. Nail, The Physical State of Mannitol after Freeze-Drying: Effects of Mannitol
445 Concentration, Freezing Rate, and a Noncrystallizing Cosolute. J. Pharm. Sci. 87 (1998) 931–935.
446 27. V.A. Drebushchak, A.G. Ogienko, E.V. Boldyreva, Polymorphic effects at the eutectic melting in the
447 H2O–glycine system. J. Therm. Anal. Calorim. 111 (2013) 2187–2194.
448 28. L.M. Lewis, R.E. Johnson, M.E. Oldroyd, Characterizing the freeze-drying behavior of model protein
449 formulations. AAPS PharmSciTech. 11 (2010) 1580–1590.
450 29. D.E. Overcashier, T.W. Patapoff, C.C. Hsu, Lyophilization of protein formulations in vials: Investigation
451 of the relationship between resistance to vapor flow during primary drying and small-scale product
452 collapse. J. Pharm. Sci. 88 (1999) 688–695.
453 30. A. Parker, S. Rigby-singleton, M. Perkins, D. Bates, D. Roux, C.J. Roberts, C. Madden-smith, L.
454 Lewis, D.L. Teagarden, R.E. Johnson, S.S. Ahmed, Determination of the Influence of Primary Drying
455 Rates on the Microscale Structural Attributes and Physicochemical Properties of Protein Containing
456 Lyophilized Products. J. Pharm. Sci. 99 (2010) 4616–4629.
457 31. M.J. Pikal, S. Shah, The collapse temperature in freeze-drying: dependence on measurement methodology
458 and rate of water removal from the glassy phase, Int. J. Pharm. 62 (1990) 165–186.
459 32. Y. Mehmood, U. Farooq, Excipients use in parenteral and lyophilized formulation development. Open Sci.
460 J. Pharm. Pharmacol. 3 (2015), 19–27.
461 33. M. Bjelošević, K.B. Seljak, U. Trestenjak, M. Logar, B. Brus, P.A. Grabnar, Aggressive condition during
462 primary drying as a contemporary approach to optimize freeze-drying cycles of biopharmaceuticals. Eur. J.
463 Pharm. Sci. 122 (2018) 292–302.
464 34. E. Meister, S. Šaši, H, Gieseler, Freeze-Dry Microscopy: Impact of Nucleation Temperature and
465 Excipient Concentration on Collapse Temperature Data. AAPS PharmSciTech. 10 (2009) 582–588

of
466 35. P. Ray, C.D. Rielly, A.G.F. Stapley, A freeze-drying microscopy study of the kinetics of sublimation in a
467 model lactose system. Chem. Eng. Sci. 172 (2017) 731–743.

ro
468 36. M.J. Pikal, S. Shah, D. Senior, J.E. Lang, Physical chemistry of freeze-drying: measurement of
469
470
-p
sublimation rates for frozen aqueous solutions by a microbalance technique. J. Pharm. Sci. 72 (1983)
635–650.
re
471 37. W.Y. Kuu, L.M. Hardwick, M.J. Akers, Rapid determination of dry layer mass transfer resistance for
472 various pharmaceutical formulations during primary drying using product temperature profiles. Int. J.
lP

473 Pharm. 313 (2006) 99–113.

474 38. E. Meister, H. Gieseler, Freeze-dry microscopy of protein/sugar mixtures: drying behavior, interpretation
na

475 of collapse temperatures and a comparison to corresponding glass transition data. J. Am. Pharm. Assoc.
476 98 (2009) 3072–3087.
ur

477 39. R.A. Depaz, S. Pansare, S.M. Patel, Freeze-drying above the glass transition temperature in amorphous
478 protein formulations while maintaining product quality and improving process efficiency. J. Pharm. Sci.
Jo

479 105 (2015) 40–49.

480 40. A. Baheti, L. Kumar, K. Arvind, A.K. Bansa, Excipients used in lyophilization of small molecules. J. Excip.
481 Food Chem. 1 (2010) 41–54.
482 41. V. Gervasi, R.D. Agnol, S. Cullen, T. McCoy, S. Vucen, A. Crean, Parenteral protein formulations: an
483 overview of approved products within the European Union. Eur. J. Pharm. Biopharm. 131 (2018) 8–24.
484 42. A. Hussein, H. Gieseler, The effect of mannitol crystallization in mannitol-sucrose systems on LDH
485 stability during freeze-drying. J. Pharm. Sci. 101 (2012) 2534–2544.
486 43. R.E. Johnson, C.F. Kirchhoff, H.T. Gaud, Mannitol-sucrose mixtures-Versatile formulation for protein
487 lyophilization. J. Pharm. Sci. 91 (2002) 914–922.
488 44. K. Greco, M. Mujat, K.L. Galbally-Kinney, D.X. Hammer, R.D. Ferguson, N. Iftimia, P. Mulhall, P.
489 Sharma, W.J. Kessler, M.J. Pikal, Accurate prediction of collapse temperature using optical coherence
490 tomography-based freeze-drying microscopy. J. Pharm. Sci. 102 (2013) 1773–1785.
491 45. G. Yang, K. Gilstrap, A. Zhang, L.X. Xu, X. He, Collapse temperature of solutions important for
492 lyopreservation of living cells at ambient temperature. Biotechnol. Bioeng. 106 (2010) 247–259.

of
ro
-p
re
lP
na
ur
Jo
 Table 1
Overview of formulations and composition of formulations in the present study.
Composition
Formulation ID Man Gly Dex Suc Tre Lac BSA
(mg/mL) (mg/mL) (mg/mL) (mg/mL) (mg/mL) (mg/mL) (mg/mL)
Man 50 formulation 50 - - - - - -
Man 100 formulation 100 - - - - - -
Man 150 formulation 150 - - - - - -
Gly 50 formulation - 50 - - - - -
Gly 100 formulation - 100 - - - - -
Gly 150 formulation - 150 - - - - -

of
Dex 50 formulation - - 50 - - - -
Dex 100 formulation - - 100 - - - -

ro
Dex 150 formulation - - 150 - - - -
Suc 50 formulation - - - 50 - - -
Suc 100 formulation
Suc 150 formulation
-
-
-
-
-
-
-p 100
150
-
-
-
-
-
-
re
Tre 50 formulation - - - - 50 - -
lP

Tre 100 formulation - - - - 100 - -

Tre 150 formulation - - - - 150 - -
Lac 50 formulation - - - - - 50 -
na

Lac 100 formulation - - - - - 100 -

Lac 150 formulation - - - - - 150 -
ur

a
Pro 0 formulation - - - - 100 - 0
a
Pro 1 formulation - - - - 100 - 1
Jo

a
Pro 5 formulation - - - - 100 - 5
a
Pro 10 formulation - - - - 100 - 10
Pro 25 formulationa - - - - 100 - 25
Abbreviations: BSA, bovine serum albumin; Dex, dextran; Gly, glycine; Lac, Lactose; Man, mannitol; Suc, sucrose;
Tre, Trehalose.
a
Protein formulation also contains 100 mM L-arginine and 0.1% polysorbate and was formulated in pH 5.5 citrate
buffer.
493
Table 2
Critical temperatures for each bulking agent.
Composition
Bulking agent
Teu (°C) Tg’ (°C) Tc (°C) Reference
Mannitol -1.4 N/A N/A [26]
Glycine -3.6 N/A N/A [27]
Dextran N/A -10 -9 [32]
a
Sucrose N/A -33.5 -32 [16], [44]
b
Trehalose N/A -30.6 -29.7 [45]
c
Lactose N/A -29.1 Approx. -27.5 [35]

of
Abbreviations: N/A, not available; Te, eutectic temperature; Tg’, glass transition temperature of maximally frozen
solution; Tc, collapse temperature measured by light transmission freeze-drying microscopy (LT-FDM).

ro
a
Tc was measure at 5% of sucrose concentration.
b
Tc was measure at 0.2 M of trehalose concentration.
c
Tc was measure at 10%-40% of lactose concentration. -p
re
lP
na
ur
Jo
494 FIGURE CAPTIONS

495 Fig. 1. Photograph of lyophilized cakes prepared by the fast and slow ramp cycles. A, Lyophilized cake
496 prepared by the fast ramp cycle at 5wt%; B, lyophilized cake prepared by the slow ramp cycle at 5wt%; C,
497 lyophilized cake prepared by the fast ramp cycle at 10wt%; D, lyophilized cake prepared by the slow ramp
498 cycle at 10wt%; E, lyophilized cake prepared by the fast ramp cycle at 15wt%; F, lyophilized cake prepared by
499 the slow ramp cycle at 15wt%.
500 Fig. 2. Defect rate (n = 6-22) of the lyophilized cake prepared by the fast and slow ramp cycles as a function of
501 the concentration of the bulking agent. A, Man formulations; B, Gly formulations; C, Dex formulations; D, Suc
502 formulations; E, Tre formulations; F, Lac formulations.

of
503 Fig. 3. Cross-sectional appearance of lyophilized cakes of Man, Dex Suc, Tre and Lac formulations. A,
504 Lyophilized cake prepared by the fast ramp cycle at 5wt%; B, lyophilized cake prepared by the fast ramp cycle

ro
505 at 15wt%; C, lyophilized cake prepared by the slow ramp cycle at 5wt%; D, lyophilized cake prepared by the
506 slow ramp cycle at 15wt%. The lyophilized cakes of Gly formulation could not be removed from the glass vials
507 while still maintaining the cake structure.
-p
re
508 Fig. 4. Microstructure of lyophilized cakes observed by scanning electron microscopy. A, Man 100 formulation
509 lyophilized by the fast ramp cycle; B, Man 100 formulation lyophilized by the slow ramp cycle; C, Dex 100
lP

510 formulation lyophilized by the fast ramp cycle; D, Dex 100 formulation lyophilized by the slow ramp cycle; E,
511 Suc 100 formulation lyophilized by the fast ramp cycle; F, Suc 100 formulation lyophilized by the slow ramp
na

512 cycle. The lyophilized cakes of the Gly formulation could not be removed from the glass vials while still
513 maintaining the cake structure.
ur

514 Fig. 5. Residual moisture of the lyophilized cake lyophilized by the fast and slow ramp cycles as a function of
515 the concentration of the bulking agent. A, Man formulations; B, Gly formulations; C, Dex formulations; D, Suc
Jo

516 formulations; E, Tre formulations; F, Lac formulations. The results are shown as the mean ± S.D (n = 3).
517 Fig. 6. Reconstitution time of the lyophilized cake lyophilized by the fast and slow ramp cycles as a function of
518 the concentration of the bulking agent. A, Man formulations; B, Gly formulations; C, Dex formulations; D, Suc
519 formulations; E, Tre formulations; F, Lac formulations. The results are shown as the mean ± S.D (n = 3).
520 Fig. 7. X-ray powder diffractogram of the lyophilized cakes. A, Man 100 formulation; B, Gly 100 formulation;
521 C, Suc 100 formulation.
522 Fig. 8. Photograph and cross-sectional appearance of Pro formulations prepared by the fast and slow ramp
523 cycles. A, Macroscopic appearance of Pro 5 formulation; B, Pro 5 formulation lyophilized by the fast ramp
524 cycle; C, Pro 25 formulation lyophilized by the fast ramp cycle; D, macroscopic appearance of Pro 25
525 formulation; E, Pro 5 formulation lyophilized by the slow ramp cycle; F, Pro 25 formulation lyophilized by the
526 slow ramp cycle.
527 Fig. 9. Defect rate, residual moisture and reconstitution time for Pro formulations. A, The results of defect rate
528 (n = 44-61) as a function of BSA concentration; B, the results of residual moisture as a function of BSA
529 concentration; C, the results of reconstitution time as a function of BSA concentration. The results for residual
530 moisture and reconstitution time are shown as the mean ± S.D (n = 3).
531 Fig. 10. Schematic illustration of the relationship between the product temperature (Tp) and the temperature to
532 initiate viscous flow during the shelf ramp process (TCSR). Case 1, expected behaviors of Tp in the further
533 aggressive cycle with fast and slow ramp rate; Case 2, expected behaviors of Tp in the aggressive cycle with fast
534 and slow ramp rate; Case 3, expected behaviors of Tp in the conservative cycle with fast and slow ramp rate.
535 Solid red line, solid blue line and black dotted line indicates shelf temperature (Ts), Tp and TCSR, respectively.

of
ro
-p
re
lP
na
ur
Jo
537
536
Jo
ur
na
lP
re
-p
ro
of
538
Jo
ur
na
lP
re
-p
ro
of
540
539
Jo
ur
na
lP
re
-p
ro
of
542
541
Jo
ur
na
lP
re
-p
ro
of
544
543
Jo
ur
na
lP
re
-p
ro
of
 545
Jo
ur
na
lP
re
-p
ro
of
Author Statement

Ryo Ohori: Conceptualization, Methodology, Validation, Formal analysis,

Investigation, Writing - original draft, Writing - review & editing, Visualization.
Tomomi Akita: Writing - review & editing. Chikamasa Yamashita: Writing - review &
editing, Supervision, Project administration.

of
ro
-p
re
lP
na
ur
Jo
Declaration of interests

■The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests:

of
ro
-p
re
lP
na
ur
Jo