Subscriber access provided by CARLETON UNIVERSITY
Article
Physiologically Based Absorption Modeling for Amorphous Solid Dispersion Formulations Amitava Mitra, Wei Zhu, and Filippos Kesisoglou Mol. Pharmaceutics, Just Accepted Manuscript • DOI: 10.1021/acs.molpharmaceut.6b00424 • Publication Date (Web): 21 Jul 2016 Downloaded from http://pubs.acs.org on July 23, 2016
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Molecular Pharmaceutics is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1
Molecular Pharmaceutics
Physiologically Based Absorption Modeling for Amorphous Solid Dispersion Formulations
2
3
Amitava Mitra*#, Wei Zhu#, Filippos Kesisoglou#
4
Biopharmaceutics
5
Pharmaceutical Sciences and Clinical Supply, Merck & Co. Inc. #
6
All authors had equal contribution
7 8 9 10 11 12
*Corresponding Author
13
Amitava Mitra, PhD
14
Merck & Co., Inc
15
West Point, PA-19486, USA
16
Tel.: +1 215 652 8551
17
Fax: +1 215 993 1245
18
E-mail:
[email protected]
19
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
20
Abstract
21
Amorphous solid dispersion (ASD) formulations are routinely used to enable the delivery of
22
poorly soluble compounds. This type of formulations can enhance bioavailability due to higher
23
kinetic solubility of the drug substance and increased dissolution rate of the formulation, by the
24
virtue of the fact that the drug molecule exists in the formulation in a high energy amorphous
25
state. In this paper we report the application of physiologically based absorption models to
26
mechanistically understand the clinical pharmacokinetics of solid dispersion formulations. Three
27
case studies are shown here to cover a wide range of ASD bioperformance in human and
28
modeling to retrospectively understand their in-vivo behavior. Case study 1 is an example of
29
fairly linear PK observed with dose escalation and the use of amorphous solubility to predict
30
bioperformance. Case study 2 demonstrates the development of a model that was able to
31
accurately predict the decrease in fraction absorbed (%Fa) with dose escalation thus
32
demonstrating that such model can be used to predict the clinical bioperformance in the scenario
33
where saturation of absorption is observed. Finally, case study 3 shows the development of an
34
absorption model with the intent to describe the observed incomplete and low absorption in
35
clinic with dose escalation. These case studies highlight the utility of physiologically based
36
absorption modeling in gaining a thorough understanding of ASD performance and the critical
37
factors impacting performance to drive design of a robust drug product that would deliver the
38
optimal benefit to the patients.
39 40 41
Keywords - Absorption Modeling, PBPK, Pharmacokinetics, Solid Dispersion, Amorphous, Dissolution
42
ACS Paragon Plus Environment
Page 2 of 38
Page 3 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
43
INTRODUCTION
44
The prediction of clinical performance of the formulations, based on preclinical data, is an
45
integral component of formulation development and a key deliverable for biopharmaceutics
46
scientists. In the last decade, physiologically based absorption modeling has attracted attention as
47
a means to achieve this preclinical to clinical translation. (1, 2, 3) Along with the advancement of
48
dissolution methods towards biorelevant measurements (4), efforts have been undertaken in the
49
field of physiologically based pharmacokinetic (PBPK) modeling to enable incorporation of
50
physicochemical and formulation characterization information into the models to simulate
51
absorption of drug compounds. (1) The principles of oral absorption PBPK models have been
52
reviewed in the literature along with a description of some of the models used in some of the
53
most common software such as such as GastroPlus (2), simCYP (3), PK-Sim (5) and
54
Intellipharm (6).
55
A review of the published literature on PBPK models for characterization or prediction of
56
formulation performance shows that the majority of publications have focused on immediate
57
release dosage forms or modified release (MR) dosage forms containing crystalline form of the
58
drug substance. For the former, a typical strategy is to describe the dissolution data using models
59
such as the Noyes-Whitney equation, and subsequent incorporation of the relevant information
60
into the absorption models. For example, a good number of publications have demonstrated the
61
successful application of these models to study the effect of particle size (7-10) or the impact of
62
dissolution rate as measured in-vitro (11-13), on absorption of these “conventional” dosage
63
forms. Similarly for MR formulations, given that the release rate from the dosage form is
64
typically the rate limiting step to absorption, direct incorporation of the measured in-vitro release
65
rates in the absorption model is employed. (10, 14-15) In addition, the development of more
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
66
robust dissolution models that accurately predict dissolution rates of low solubility molecules
67
have been recently reported. (16)
68
With the increase in the number of insoluble compounds in drug development pipelines (17),
69
often new formulation approaches are employed to achieve adequate bioavailability. Such
70
approaches may include the utilization of nanosuspensions, lipids or co-solvents, and amorphous
71
solid dispersions among others. (18) Among these approaches, amorphous solid dispersions
72
(ASD) have recently received increased attention with several new drugs on the market. (19, 20)
73
Amorphous solid dispersion (ASD) holds the promise of superior performance because of the
74
inherently higher energy state of active pharmaceutical ingredients (API) over the crystalline
75
counterpart. (21) The ASDs typically enhance bioavailability due to higher kinetic solubility of
76
the drug substance and increased dissolution rate of the formulation, by the virtue of the fact that
77
the drug molecule exists in the formulation in a high energy amorphous state. However, a higher
78
energy alone does not guarantee higher bioavailability. Amorphous materials have unique
79
challenges as compared to other more common approaches for increasing bioavailability due to
80
an increased risk of crystallization to a more stable crystalline form. Thus it would appear
81
beneficial to expand the application of physiologically based absorption modeling to these
82
formulations. However, there are not many published reports available on simulation of
83
absorption from these “enabled” dosage forms and from solid dispersions in particular. While the
84
principles of absorption from supersaturated solutions have been studied (22), publications on
85
applications of such models during product development are very limited. Gao et al. (23)
86
described the prediction of performance of a solid dispersion, among other formulations, for rat
87
toxicology studies while more recently Zheng et al, (24) discussed the utility of PBPK models in
88
the discovery space to drive utilization of a solid dispersion. To the best of our knowledge, there
ACS Paragon Plus Environment
Page 4 of 38
Page 5 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
89
have been no detailed reports of clinical data from solid dispersion formulations and application
90
of absorption models to describe the clinical data.
91
In this manuscript, we report three case studies attempting to highlight successes and challenges
92
in utilizing physiologically based absorption models for solid dispersion formulations. Given the
93
complexity of the dissolution behavior of solid dispersions (25), we focused on formulations that
94
could be categorized as supersaturating systems without evidence of complicated species
95
formation such as nanoparticles in in-vitro dissolution studies. The compounds exhibited varied
96
absorption and pharmacokinetic behavior ranging from complete absorption with linear
97
pharmacokinetics for case study 1 to incomplete and low absorption for case study 3. We discuss
98
the development of models either based on the in-vitro solubility and dissolution data or based
99
on extrapolation of models developed to first describe formulation performance in a preclinical
100
dog model. Extensive physical characterization of these ASDs was carried out to ensure that they
101
maintained their amorphous nature prior to in-vitro and in-vivo studies. However, discussion of
102
those data is outside the scope of this paper, since the focus here is on development of absorption
103
models.
104 105
MATERIALS and METHODS
106
Software Used
107
GastroPlus v 8.5 (Simulations Plus, Lancaster, CA, USA) was used for absorption modeling in
108
all the case studies.
109 110 111
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
112
Case Study 1: Linear pharmacokinetics with high fraction absorbed
113
Merck compound A physicochemical properties: Compound A is a BCS class II free base
114
molecule. The following key compound properties were used in building the model – molecular
115
weight 442, log P 4.21, pKa 3.53, density 1.2 g/mL, calculated human effective permeability
116
(based on Caco-2 data) was 3.2 x 10-4 cm/sec and diffusion coefficient was calculated based on
117
molecular weight. The amorphous solubility values from dissolution study were input as pH
118
solubility profile, as described below. The solubility of the crystalline form in simulated gastric
119
fluid (SGF) and fasted state simulated intestinal fluid (FaSSIF) was 0.02 mg/mL and 0.004
120
mg/mL, respectively. Diffusion coefficient was adjusted in GastroPlus to account for bile salt
121
concentration changes across the gastrointestinal (GI) tract based on diffusion coefficient and in-
122
vivo dissolution rate. The precipitation time was fixed at the default value of 900 seconds. The
123
blood to plasma ratio was 0.83. Fraction unbound in plasma was 0.9%.
124
Dosage form and dissolution data input: The compound was formulated as an amorphous solid
125
dispersion using HPMC-AS as the polymer matrix. Immediate release tablet was chosen as the
126
formulation option in GastroPlus. The dissolution data for this formulation was generated in 2-
127
stage dissolution (SGF followed by FaSSIF). The samples from the dissolution studies were
128
ultracentrifuged at 80,000 rpm before analysis to remove any undissolved API. Furthermore the
129
dissolution data did not show any indication of crystallization during the time-frame of the study.
130
Hence this data represented amorphous solubility of the compound. This dissolution data was
131
input in the model as a pH- amorphous solubility profile – 0.43 mg/mL (pH 1.3), 0.2 mg/mL (pH
132
6.0), 0.013 mg/mL (pH 6.2), and 0.015 mg/mL (pH 6.5). The solubility profile was input to
133
maintain supersaturation in the small in the small intestine, as there were no signs of
134
crystallization in the dissolution study. Based on particle size measurement of the dissolving
ACS Paragon Plus Environment
Page 6 of 38
Page 7 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
135
solid dispersion in the dissolution study, the mean size of the dissolving particle was set to be 1
136
µm (diameter) in the model. Similar particle size was assumed for the crystalline form of the
137
API, so that there is direct comparison of the effect of solubility difference of the API forms on
138
PK.
139
Physiology: The default human fasted physiological model in GastroPlus (Opt logD SA/v6.1)
140
was used in these simulations.
141
Pharmacokinetic (PK) parameters: Human PK parameters were estimated by fitting the 2 mg oral
142
data from first-in-human (FIH) study (Merck data on file) to a two-compartment model in PK-
143
plus. The mean PK parameters used in these simulations were CL/F = 0.235 L/hr/kg and V/F =
144
0.398 L/kg, k12 = 0.653 hr-1, and k21 = 0.25 hr-1. Based on the low dose and linear
145
pharmacokinetics it was assumed that F=1 (i.e. CL/F and V/F were used directly to simulate
146
systemic disposition)
147
Simulation: Single simulations were conducted to predict the mean PK profiles and parameters
148
for several doses from the FIH study using the amorphous solubility data from the dissolution
149
study. Additional simulations were conducted using the crystalline solubility of the compound to
150
investigate whether these solubility data were able to predict the observed FIH data. Parameter
151
sensitivity analysis (PSA) was conducted to assess the impact of precipitation time (90 – 9000
152
seconds) on fraction absorbed as function of dose (2 – 400 mg).
153 154
Case Study 2: Saturation of absorption
155
Merck compound B physicochemical properties: Compound B is a BCS class II molecule. The
156
following key compound properties were used in building the model – molecular weight 624.8,
157
log D (pH 7) 2.63, density 1.2 g/mL, calculated human effective permeability (based on Caco-2
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
158
data) was 2.38 x 10-4 cm/sec and diffusion coefficient was calculated based on molecular weight.
159
The solubility in buffer was 0.021 mg/mL and there was no pH dependent solubility profile. The
160
amorphous solubility in SGF and in FaSSIF was 0.021 mg/mL and 0.035 mg/mL, respectively.
161
The amorphous solubility values were obtained from the dissolution data as described in case
162
study 1. The solubility and diffusion coefficient was adjusted in GastroPlus to account for
163
changes across the gastrointestinal (GI) tract based on bile salt concentration changes, by fitting
164
the SGF and FaSSIF solubility data. The size of the dissolving particle was assumed to be 1 µm
165
(diameter), based on particle size measurement of the dissolving solid dispersion in the
166
dissolution study. The precipitation time was fixed at the default value of 900 seconds because
167
there is no pH dependent solubility of this compound and dissolution studies have shown no
168
indication of precipitation, therefore precipitation rate is expected to have minimal impact on
169
fraction absorbed. The blood to plasma ratio was set as 1. Fraction unbound in plasma was set as
170
5%.
171
Dosage form input: The compound was formulated as an amorphous solid dispersion using
172
HPMC-AS as the polymer matrix. Immediate release tablet was chosen as the formulation option
173
in GastroPlus.
174
Physiology: The default human fasted physiological model in GastroPlus (Opt logD SA/v6.1)
175
was modified by reducing absorption from caecum and ascending colon by manually reducing
176
the ASF to 1.0 in both caecum and ascending colon. This was done due to low buffer solubility
177
of the compound and no evidence of prolonged absorption in preclinical studies. Based on
178
experience with similar compounds, it was assumed that this compound would have minimal
179
absorption in the lower GI tract in human.
ACS Paragon Plus Environment
Page 8 of 38
Page 9 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
180
Pharmacokinetic (PK) parameters: Human PK parameters were estimated by fitting the 25 mg,
181
which was the starting dose, oral data from first-in-human (FIH) study (Merck data on file) to a
182
two-compartment model in WinNonin v5.3. The mean PK parameters used in these simulations
183
were CL = 0.66 L/hr/kg and V = 1.84 L/kg, k12 = 0.139 hr-1, and k21 = 0.190 hr-1. Based on the
184
low starting dose, linear pharmacokinetics up to 500 mg dose and lack of food effect at 125 mg
185
dose, it was assumed that fraction absorbed (Fa) = 1 at 25 mg. The first pass extraction (FPE)
186
was back-calculated as 51% using hepatic blood flow and blood/plasma ratio.
187
Simulation: Single simulations were conducted to predict the mean PK profiles and parameters
188
for several doses from the FIH study using the amorphous solubility data from the dissolution
189
study.
190 191
Case Study 3: Incomplete and low absorption
192
Merck compound C physicochemical properties: Compound C is a BCS IV compound with low
193
solubility (~0.1 µg/mL) across the physiological pH range and moderate permeability (LLCPK1
194
Papp = 7.9 x 10-6 cm/sec). Given the very low pKa ( 90%) across all the dose levels. Some fluctuations in the PK profile in the
258
elimination phase at the higher doses were likely an artifact of the model and have no bearing on
259
the overall outcome.
260 261
To investigate whether the use of amorphous solubility was critical to be able to accurately
262
predict the observed clinical data at all the doses, additional simulations were carried out using
263
the crystalline solubility of compound A. As shown in figure 2A, the low dose (2 mg) was
264
simulated accurately. However at 50 mg and above, underprediction of AUC and Cmax was
265
observed along with a right shift in Tmax. Figure 2B shows the simulations at 200 mg. These
266
simulations clearly demonstrated that solubility of the amorphous form of the API was needed to
267
predict the observed PK in the full FIH dose range. It is possible at the intermediate and higher
268
doses some mixture of amorphous and crystalline forms co-exists. However the current version
269
of the software doesn’t have the ability to handle two solubility values simultaneously hence
270
such a scenario cannot be simulated without significant assumptions. Finally, parameter
ACS Paragon Plus Environment
Page 13 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
271
sensitivity analysis demonstrated that there was minimal impact on fraction absorbed (90-100%)
272
over a wide range of precipitation time in the FIH dose range.
273 274
Case Study 2: Saturation of absorption
275
The simulated plasma concentration vs. time profiles and the PK parameters are shown in Figure
276
3 and Table 2, respectively, for doses evaluated in FIH study. The observed data are shown for
277
comparison. Overall, the model was able to predict observed human PK data with reasonable
278
accuracy across the doses, except that delayed Tmax was predicted at 125 mg dose and above,
279
which was not consistent with the observed Tmax across doses.
280 281
Figure 4 depicts the observed dose normalized AUC and simulated fraction absorbed (Fa) for
282
compound B at individual doses. The trend of saturation of absorption was observed in human
283
PK data, as dose normalized AUC0-t decreases from 0.108 to 0.037 µg/mL*hr/mg with increase
284
in dose. Similarly, a 3~4 fold decrease in Fa (%) was predicted in the absorption model across
285
the doses from 25 mg to 750 mg. This demonstrated that the model was able to capture the
286
observed saturation of absorption with reasonable accuracy.
287 288
Case Study 3: Incomplete and low absorption
289
The results of the dog pharmacokinetic study at 2 mg/kg dose and the modeling of the data are
290
shown in Figure 5. The overall bioavailability was low (approximately 10-15%). As shown in
291
Figure 5 an initial simulation based on the amorphous API solubility resulted in significant
292
underprediction of the data. An increase in solubility by 4-fold (implemented in the software by
293
increasing the bile salt solubilization ratio (SR); very comparable results are obtained by
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 38
294
adjusting the baseline solubility and maintaining the initial SR value) resulted in a better
295
description of the data; the overall extent of absorption was reasonably well captured although
296
the absorption phase was not fully captured. Since the intent of this dog simulation was not to
297
fully describe the dog data but to get an approximation of in vivo solubility with minimal model
298
changes, this solubility estimate was considered acceptable in an early development setting for
299
human PK projections. It is likely that the permeability in the dog was somewhat underpredicted
300
from the current model. However since there is no direct translation of dog to human
301
permeability that aspect of the model was not modified. This 4-fold higher in-vivo solubility
302
was subsequently used in the human simulations, with the assumption that the higher in-vivo
303
solubility observed in dogs is representative of human in-vivo solubility. It should be noted that
304
compound C plasma concentration at 24 hours post dose was unexpectedly high, but such
305
observations have been reported previously. (27) It is acknowledged that an increase in solubility
306
due to better bile salt solubilization in vivo may not represent the only solution to obtaining an
307
improved data fit. Thus additional simulations were conducted exploring alternative permeability
308
values or different particle size. An effective permeability of 6.7 × 10-4 cm/sec provided a similar
309
fit as that to the increase of in-vivo solubility (data not shown). This is a very high permeability
310
value (only a couple of compounds such as ketoprofen and piroxicam have been reported to have
311
such high permeability), that would appear unlikely given the cell line permeability and the
312
molecular weight (~600 Da) of compound C. Simulations conducted using small in vivo particle
313
size could not account for the observed exposures. Even when 0.2 µm (diameter) particle size
314
was used for the simulation, plasma concentrations were significantly underpredicted
315
(approximately 3-fold), similar to the simulations with the original solubility.
316
ACS Paragon Plus Environment
Page 15 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
317
Figure 6 depicts the simulation of clinical data at doses of 5, 15, 50 and 150 mg. Higher doses
318
(400 mg and 800 mg) are not shown as will be explained below. While simulations based on the
319
amorphous API solubility accurately predicted the plasma concentration vs time profile for the
320
first two doses (5 and 15 mg – Figure 6 panels A and B), they significantly underpredicted the
321
pharmacokinetics at the higher doses. Given the significant underprediction at 150 mg (Figure 6
322
panel D), simulations of higher doses are not shown. Incorporation of the in-vivo solubility
323
estimate from the dog model (i.e. 4 fold higher in-vivo solubility) significantly improved the
324
prediction at the 150 mg dose, which was a similar dose to that used in the dog study (2 mg/kg).
325
Although similar to the simulations of the dog data some underprediction of the absorption phase
326
was observed.
327 328
DISCUSSION
329
ASDs enable higher bioavailability as compared to formulation made from crystalline drug
330
substance due to higher solubility and dissolution rate of the amorphous product. However, it has
331
been observed that rendering a material amorphous does not guarantee success. (28) Hence
332
mechanistic understanding of the factors providing enhanced bioperformance (or lack of) is
333
imperative. PBPK modeling to prospectively and/or retrospectively understand the absorption
334
and bioperformance of ASDs in clinic based on in-vitro dissolution data and/or preclinical PK
335
data could be very important to enable successful development of these formulations. The three
336
case studies shown here were chosen to cover a wide range of ASD bioperformance in human
337
and modeling to retrospectively understand their in-vivo behavior. Case study 1 is an example of
338
fairly linear PK observed in the FIH study and the use of amorphous solubility to predict
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
339
bioperformance, case study 2 exemplifies saturation of absorption observed in FIH and case
340
study 3 demonstrates a formulation showing in-complete and low absorption.
Page 16 of 38
341 342
Compound A (case study 1) was formulated as a spray dried dispersion using HPMC-AS as the
343
polymer matrix. In single ascending dose FIH study, reasonable dose proportionality in AUC
344
was observed throughout the dose range of 2-400 mg but Cmax showed less than dose
345
proportionality at the highest dose (400 mg). Dissolution of this formulation in SGF followed by
346
FaSSIF, showed that in FaSSIF the amorphous formulation was able to provide approximately 4-
347
fold higher solubility as compared to the crystalline form of the API and also this supersaturation
348
was maintained throughout the duration (120 minutes) of the dissolution experiment. An
349
absorption model built using these amorphous solubility data from the dissolution experiment
350
was able to predict the PK profile (Figure 1) and parameters (Table 1) with reasonable accuracy.
351
Thus demonstrating that in-vivo the formulation was able to achieve similar amorphous
352
solubility as shown in the in-vitro dissolution experiment and also maintain supersaturation
353
throughout the dose levels. This was further confirmed by the results of simulations conducted
354
using solubility of the crystalline API, which showed that while at low dose (2 mg) the model
355
predicted the observed data well (Figure 2A), there was significant underprediction at 200 mg
356
(Figure 2B). Thus demonstrating that the formulation was able to maintain amorphous solubility
357
in-vivo at all doses in-order to achieve the observed dose proportional exposure. Since
358
compound A is a weakly basic amorphous compound, precipitation in the GI tract might impact
359
its bioavailability. However, sensitivity analysis (Figure 2C) showed that precipitation time had
360
minimal impact on fraction absorbed. This is likely because the amorphous solubility of the
361
compound is high enough across the GI tract that it maintains supersaturation throughout the
ACS Paragon Plus Environment
Page 17 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
362
dose range studied in the clinic. These simulations clearly demonstrate that a thorough
363
understanding of behavior of these amorphous formulations in appropriately designed dissolution
364
experiments and subsequently incorporating that data in the PBPK model is needed to be able to
365
predict their bioperformance.
366 367
Compound B (case study 2) was formulated as a spray dried amorphous solid dispersion using
368
HPMC-AS as the polymer matrix. In single ascending dose FIH study, both AUC and Cmax
369
showed less than dose proportionality PK as the dose increases from 25 mg to 750 mg, indicating
370
saturation of absorption at higher doses. Dissolution of this formulation showed that the
371
dissolved drug of solid dispersion formulation was able to provide approximately 2-fold higher
372
solubility as compared to a formulation with crystalline form of the API, and the supersaturation
373
was maintained throughout the duration (90 minutes) of the dissolution experiment in FaSSIF.
374
An absorption model was built using the amorphous solubility data from the dissolution
375
experiment to predict the clinical PK profiles (Figure 3) and PK parameters (Table 2) with
376
reasonable accuracy, although a somewhat delayed Tmax was predicted especially at higher doses.
377
The reason for lack of prediction of the absorption phase at higher doses could be due to factors
378
such as inadequate estimation of in-vivo solubility even with use of bile salt based solubilization
379
or inadequate effective permeability estimation or a combination of these factors. To further
380
illustrate the saturation of absorption in FIH study, the observed dose normalized AUC and
381
simulated %Fa for compound B at individual doses were plotted in Figure 4, which showed that
382
the degree of saturation of absorption in observed human PK data are in agreement with the
383
decrease of %Fa predicted by the absorption model across the doses from 25 mg to 750 mg. The
384
less than dose proportional PK could be an indication that the amorphous solubility limit has
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 18 of 38
385
been reached at the higher doses resulting in a solubility-limited absorption. In addition, it is
386
possible that the supersaturation achieved in-vivo in the GI fluid may not be maintained as well
387
as that suggested by the in-vitro dissolution system, which could also contribute to a less-than-
388
dose proportional increase in PK. While it is acknowledged that in this particular case the Tmax
389
was not precisely captured, the model was able to accurately predict the decrease in fraction
390
absorbed with dose escalation. Thus demonstrating that such physiologically based oral
391
absorption model can be used to model the clinical bioperformance in the scenario where
392
saturation of absorption is observed.
393 394
Compound C (case study 3) represents an interesting and challenging case study for oral
395
absorption modeling. Based on our experience, the behavior of compound C would have been
396
difficult to model with the currently available absorption models even if this formulation was a
397
simple crystalline API suspension. Specifically, compound C appears to exhibit relatively low
398
bioavailability (based on preclinical estimates and experience with other compounds in the
399
series) across all doses, however saturation of absorption is not seen until relatively high dose
400
(400 mg and higher). This relatively constant bioavailability across a wide dose range 5-400 mg
401
is very difficult to model based on the known compound and formulation properties. It is likely
402
that the low permeability of the compound is playing a role in the behavior and may be coupled
403
with regiodependent absorption (no data were available during development of Compound C to
404
assess this possibility). It is generally acknowledged that BCS IV compounds are the most
405
difficult to handle from absorption modeling standpoint. (1, 29)
406
ACS Paragon Plus Environment
Page 19 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
407
The better fit at 150 mg (Figure 6, panel D) based on extrapolated in-vivo solubility from the dog
408
data suggests that this preclinical to clinical translation may be one way to inform simulations for
409
these challenging compounds. Others have also suggested this preclinical to clinical translation
410
step as an initial step in absorption modeling. Recently Zhang et al. used estimates of in-vivo
411
solubilities from rats and dogs to successfully predict FIH pharmacokinetics for eighteen
412
compounds. (30) However one limitation with this approach is that extrapolation to doses outside
413
of those that have been tested in preclinical species may not be straightforward. It is also possible
414
that for solid dispersions (and other solubilizing formulations) the dose to excipient ratio also
415
dictates in-vivo solubilization capacity (or crystallization potential), which can make preclinical
416
to human translation complicated. In the case of compound C, the extrapolation at a comparable
417
mg/kg dose (where dose/excipient/dosing volume ratios were quite similar) appeared to work,
418
although based on the simulations shown in Figure 6 it is evident that the same higher solubility
419
would overpredict exposures at lower doses (data not shown). In a typical drug/formulation
420
development setting, prior to FIH studies would be conducted at a dose representing the
421
projected clinical dose and/or the upper range of the FIH study. This was the case for compound
422
C. While this approach works to project the human PK at the intended dose, it is acknowledged
423
that more studies may be required to project the behavior across the entire FIH dose range (if that
424
is of interest). The final projection at the clinical dose is within 2-fold of the observed data,
425
which would be considered a generally acceptable prediction error for these types of models at
426
early development stage.
427 428 429
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 20 of 38
430
CONCLUSION
431
For amorphous solid dispersion formulations understanding the factors that impact their
432
bioperformance is critical because by design these formulations are thermodynamically unstable
433
with propensity to crystalize and they have complex dissolution characteristics. Physiologically
434
based oral absorption modeling can play an integral part in advancing biopharmaceutics
435
knowledge of these formulations and help in robust design of formulation per the quality by
436
design (QbD) paradigm. The case studies discussed in this manuscript demonstrate the potential
437
application of absorption modeling in such context, with three examples showing increasing
438
levels of biopharmaceutic complexity. Case study 1 is an example of successful application of
439
amorphous solubility obtained from dissolution experiment to predict the linear PK observed
440
with dose escalation in the FIH study. Case study 2 demonstrates that these models can be used
441
to simulate the clinical bioperformance in the scenario where saturation of absorption is observed
442
with dose escalation. Finally, case study 3 exemplifies the application of such model to describe
443
the observed in-complete and low absorption with dose escalation. The biopharmaceutics
444
knowledge gained from these absorption models would maximizes the mechanistic
445
understanding of critical factors affecting the bioperformance of solid dispersion formulations
446
and thereby help in development of a final product that would deliver the optimal benefit to the
447
patients.
448 449
ACKNOWLEDGEMENTS
450
The authors would like to thank the project teams for providing the critical data needed to build
451
the models.
452
ACS Paragon Plus Environment
Page 21 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
453 454
Molecular Pharmaceutics
REFERENCES 1. Kostewicz, E.S.; Aarons, L.; Bergstrand, M.; Bolger, M.B.; Galetin, A.; Hatley, O.;
455
Jamei, M.; Lloyd, R.; Pepin, X.; Rostami-Hodjegan, A.; Sjögren, E.; Tannergren, C.;
456
Turner, D.B.; Wagner, C.; Weitschies, W.; Dressman, J. PBPK Models for the Prediction
457
of in vivo Performance of Oral Dosage Forms. Eur J Pharm Sci. 2014, 57, 300-321.
458
2. Agoram, B.; Woltosz, W.S.; Bolger, M.B. Predicting the Impact of Physiological and
459
Biochemical Processes on Oral Drug Bioavailability. Adv Drug Deliv Rev. 2001, 50
460
Suppl 1, S41-67.
461
3. Jamei, M.; Turner, D.; Yang, J.; Neuhoff, S.; Polak, S.; Rostami-Hodjegan, A.; Tucker,
462
G. Population-based Mechanistic Prediction of Oral Drug Absorption. AAPS J. 2009, 11,
463
225-237.
464
4. Kostewicz, E.S.; Abrahamsson, B.; Brewster, M.; Brouwers, J.; Butler, J.; Carlert, S.;
465
Dickinson, P.A.; Dressman, J.; Holm, R.; Klein, S.; Mann, J.; McAllister, M.; Minekus,
466
M.; Muenster, U.; Müllertz, A.; Verwei, M.; Vertzoni, M.; Weitschies, W.; Augustijns, P.
467
In vitro Models for the Prediction of in vivo Performance of Oral Dosage Forms. Eur J
468
Pharm Sci. 2014, 57, 342-366.
469
5. Thelen, K.; Coboeken, K.; Willmann, S.; Dressman, J.B.; Lippert, J. Evolution of a
470
Detailed Physiological Model to Simulate the Gastrointestinal Transit and Absorption
471
Process in Humans, Part II: Extension to Describe Performance of Solid Dosage Forms. J
472
Pharm Sci. 2012, 101, 1267-1280.
473
6. Johnson, K.C. Dissolution and Absorption Modeling: Model Expansion to Simulate the
474
Effects of Precipitation, Water Absorption, Longitudinally Changing Intestinal
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
475
Permeability, and Controlled Release on Drug Absorption. Drug Dev Ind Pharm. 2003,
476
29, 833-842.
477 478 479 480
Page 22 of 38
7. Kesisoglou, F.; Mitra, A. Application of Absorption Modeling in Rational Design of Drug Product Under Quality-by-Design Paradigm. AAPS J. 2015, 17, 1224-1236. 8. Kesisoglou, F.; Wu, Y. Understanding the Effect of API Properties on Bioavailability Through Absorption Modeling. AAPS J. 2008, 10, 516-525.
481
9. Parrott, N.; Hainzl, D.; Scheubel, E.; Krimmer, S.; Boetsch, C.; Guerini, E.; Martin-
482
Facklam, M. Physiologically based Absorption Modelling to Predict the Impact of Drug
483
Properties on Pharmacokinetics of Bitopertin. AAPS J. 2014, 16, 1077-1084.
484
10. Kesisoglou, F.; Chung, J.; Van Asperen,
J.; Heimbach, T. Physiologically Based
485
Absorption Modeling to Impact Biopharmaceutics and Formulation Strategies in Drug
486
Development—Industry Case Studies. J Pharm Sci. 2016, (article in press)
487
11. Mitra, A.; Kesisoglou, F.; Dogterom, P. Application of Absorption Modeling to Predict
488
Bioequivalence Outcome of Two Batches of Etoricoxib Tablets. AAPS PharmSciTech.
489
2015, 16, 76-84.
490
12. Crison, J.R.; Timmins, P.; Keung, A.; Upreti, V.V.; Boulton, D.W.; Scheer, B.J.
491
Biowaiver Approach for Biopharmaceutics Classification System Class 3 Compound
492
Metformin Hydrochloride Using In Silico Modeling. J Pharm Sci. 2012, 101, 1773-1782.
493
13. Sperry, D.C.; Thomas, S.J.; Lobo, E. Dissolution Modeling of Bead Formulations and
494
Predictions of Bioequivalence for a Highly Soluble, Highly Permeable Drug. Mol Pharm.
495
2010, 7, 1450-1457.
496 497
14. Mathias, N.R.; Crison, J. The Use of Modeling Tools to Drive Efficient Oral Product Design. AAPS J. 2012, 14, 591-600.
ACS Paragon Plus Environment
Page 23 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
498
15. Chen, W.; Desai, D.; Good, D.; Crison, J.; Timmins, P.; Paruchuri, S.; Wang, J.; Ha, K.
499
Mathematical Model-Based Accelerated Development of Extended-release Metformin
500
Hydrochloride Tablet Formulation. AAPS PharmSciTech. 2015, Oct 19. [Epub ahead of
501
print]
502
16. Wang, Y.; Abrahamsson, B.; Lindfors, L.; Brasseur, J.G. Analysis of Diffusion-
503
Controlled Dissolution from Polydisperse Collections of Drug Particles with an Assessed
504
Mathematical Model. J Pharm Sci. 2015, 104, 2998-3017.
505 506
17. Venkatesh, S.; Lipper R. A. Role of the Development Scientist in Compound Lead Selection and Optimization. J Pharm Sci. 2000, 89, 145-154.
507
18. Williams, H.D.; Trevaskis, N.L.; Charman, S.A.; Shanker, R.M.; Charman, W.N.;
508
Pouton, C.W.; Porter, C.J. Strategies to Address Low Drug Solubility in Discovery and
509
Development. Pharmacol Rev. 2013, 65, 315-499.
510
19. Baghel, S.; Cathcart, H.; O'Reilly, N.J.; Polymeric Amorphous Solid Dispersions: A
511
Review of Amorphization, Crystallization, Stabilization, Solid-State Characterization,
512
and Aqueous Solubilization of Biopharmaceutical Classification System Class II Drugs. J
513
Pharm Sci. 2016 [Epub ahead of print].
514
20. Rumondor, A.C..; Dhareshwar, S.S.; Kesisoglou, F. Amorphous Solid Dispersions or
515
Prodrugs: Complementary Strategies to Increase Drug Absorption, J Pharm Sci. 2016
516
[Epub ahead of print]
517
21. Serajuddin, A.T.M. Solid dispersion of Poorly Water-Soluble Drugs: Early Promises,
518
Subsequent Problems, and Recent Breakthroughs. J Pharm Sci. 1999, 88, 1058-1066.
519
22. Miller, J.M.; Beig, A.; Carr, R.A.; Spence, J.K.; Dahan, A. A Win-Win Solution in Oral
520
Delivery of Lipophilic Drugs: Supersaturation via Amorphous Solid Dispersions
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
521
Increases Apparent Solubility without Sacrifice of Intestinal Membrane Permeability.
522
Mol Pharm. 2012, 9, 2009-2016.
Page 24 of 38
523
23. Gao, Y.; Carr, R.A.; Spence, J.K.; Wang, W.W.; Turner, T.M.; Lipari, J.M.; Miller, J.M.
524
A pH-dilution Method for Estimation of Biorelevant Drug Solubility along the
525
Gastrointestinal Tract: Application to Physiologically based Pharmacokinetic Modeling.
526
Mol Pharm. 2010, 7, 1516-1526.
527
24. Zheng, W.; Jain, A.; Papoutsakis, D.; Dannenfelser, R.M.; Panicucci, R.; Garad, S.
528
Selection of Oral Bioavailability Enhancing Formulations during Drug Discovery. Drug
529
Dev Ind Pharm. 2012, 38, 235-247.
530
25. Friesen, D.T.; Shanker, R.; Crew, M.; Smithey, D.T.; Curatolo, W.J.; Nightingale, J.A.
531
Hydroxypropyl Methylcellulose Acetate Auccinate-based Spray-Dried Dispersions: An
532
Overview. Mol Pharm. 2008, 5, 1003-1019.
533
26. Kesisoglou, F. Use of Preclinical Dog Studies and Absorption Modeling to Facilitate Late
534
Stage Formulation Bridging for a BCS II Drug Candidate. AAPS PharmSciTech. 2014,
535
15, 20-28.
536
27. Benincosa, L.J.; Audet, P.R.; Lundberg, D.; Zariffa, N.; Jorkasky, D.K. Pharmacokinetics
537
and Absolute Bioavailability of Epristeride in Healthy Male Subjects. Biopharm Drug
538
Dispos. 1996, 17, 249-258.
539 540
28. Newman, A.; Knipp, G.; Zografi, G. Assessing the Performance of Amorphous Solid Dispersions. J Pharm Sci. 2012, 101, 1355-1377.
541
29. Lennernas, H.; Aarons, L.; Augustijns, P.; Beato, S.; Bolger, M.; Box, K.; Brewster, M.;
542
Butler, J.; Dressman, J.; Holm, R.; Julia Frank, K.; Kendall, R.; Langguth, P.; Sydor, J.;
543
Lindahl, A.; McAllister, M.; Muenster, U.; Mullertz, A.; Ojala, K.; Pepin, X.; Reppas, C.;
ACS Paragon Plus Environment
Page 25 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
544
Rostami-Hodjegan, A.; Verwei, M.; Weitschies, W.; Wilson, C.; Karlsson, C.;
545
Abrahamsson, B. Oral Biopahrmaceuotcs Tools – Time for a New Initiative – An
546
Introduction to the IMI Project OrBito. Eu J Pharm Sci. 2014, 57, 292-299.
547 548
30. Zhang, T.; Heimbach, T.; Lin, W.; Zhang, J.; He, H. Prospective Predictions of Human Pharmacokinetics for Eighteen Compounds. J Pharm Sci. 2015, 104, 2795-2806.
549
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
550
TABLE LEGENDS
551
Table 1: Simulated and observed pharmacokinetic parameters for compound A at dose range of
552
2 - 400 mg.
553
Table 2: Simulated and observed pharmacokinetic parameters for compound B at dose range of
554
25 – 750 mg.
555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572
ACS Paragon Plus Environment
Page 26 of 38
Page 27 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Molecular Pharmaceutics
573
FIGURE LEGENDS
574
Figure 1: (A) Simulated and observed plasma concentration vs. time profiles for compound A at
575
2 mg, 4 mg, 8 mg and 15 mg doses. (B) Simulated and observed plasma concentration vs. time
576
profiles for compound A at 25 mg, 50 mg, 100 mg, 200 mg and 400 mg doses.
577 578
Figure 2: Simulated plasma concentration vs. time profiles for compound A at 2 mg dose (A),
579
200 mg dose (B) using solubility of the crystalline form of the compound. The observed plasma
580
concentration vs. time profiles at both doses is shown for comparison. (C) Parameter sensitivity
581
analysis showing the impact of precipitation time and dose on fraction absorbed.
582 583
Figure 3: Simulated and observed plasma concentration vs. time profiles for compound B at 25
584
mg, 50 mg, 125 mg, 250 mg, 500 mg and 750 mg doses.
585 586
Figure 4: Observed dose normalized AUC0-t (µg/mL*hr/mg) and simulated fraction absorbed
587
(%Fa) for compound B at 25 mg, 50 mg, 125 mg, 250 mg, 500 mg and 750 mg doses.
588 589
Figure 5 Simulated (lines) and observed (squares) plasma concentration vs. time profiles for
590
compound C in dogs at a 2 mg/kg dose. Initial simulation based on amorphous API solubility
591
(dashed line) and final model fit to 4-fold higher in-vivo solubility (solid line) are shown.
592 593
Figure 6 Simulated (dashed lines) and observed (squares) plasma concentration vs. time profiles
594
for compound C at 5 mg, 15 mg, 50 mg and 150 mg doses (panels A-D, respectively). For the
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
595
150 mg dose an additional simulation based on the in-vivo solubility obtained from dog data is
596
depicted (solid line).
597
ACS Paragon Plus Environment
Page 28 of 38
Page 29 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
598
Molecular Pharmaceutics
Table 1 Observed Dose (mg)
Predicted
2
AUC0-24hr (µg/mL*hr) 0.12
Cmax (µg/mL) 0.018
Tmax (hr) 1.0
AUC0-24hr (µg/mL*hr) 0.13
Cmax (µg/mL) 0.022
Tmax (hr) 1.4
4
0.25
0.037
2.0
0.23
0.047
1.4
8
0.49
0.069
2.0
0.46
0.094
1.4
15
1.03
0.163
2.0
0.86
0.177
1.4
25
1.61
0.216
2.0
1.43
0.295
1.7
50
2.89
0.551
2.0
2.86
0.559
1.7
100
6.40
1.106
2.0
5.72
0.978
2.0
200
11.5
1.92
1.0
11.4
1.79
1.0
400
19.1
1.94
2.0
20.3
2.30
1.2
599 600 601 602 603
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
604
Page 30 of 38
Table 2 Observed Dose
Predicted
AUC0-t (µg/mL*hr)
Cmax (µg/mL)
Tmax (hr)
AUC0-t (µg/mL*hr)
Cmax (µg/mL)
Tmax (hr)
Fa (%)
25
0.270
0.043
2.0
0.256
0.051
1.8
100
50
0.398
0.058
1.5
0.528
0.095
2.0
100
125
0.861
0.088
1.5
1.155
0.144
3.5
87
250
1.123
0.113
1.0
1.563
0.161
3.5
59
500
2.131
0.180
2.0
1.908
0.175
4.5
36
750
2.810
0.249
2.0
2.094
0.181
4.8
26
(mg)
605 606 607
ACS Paragon Plus Environment
Page 31 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
608
Molecular Pharmaceutics
Figure 1 (A)
609
(B)
610
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
611
Figure 2
(A)
612
(B)
613 614
ACS Paragon Plus Environment
Page 32 of 38
Page 33 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
615 616
Molecular Pharmaceutics
Figure 2
617
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
618
Figure 3
619 620
ACS Paragon Plus Environment
Page 34 of 38
Page 35 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
621
Molecular Pharmaceutics
Figure 4
622 623
ACS Paragon Plus Environment
Molecular Pharmaceutics
624
Figure 5
625 Observed data
160
Simulation based on amorphous API solubility Simulation with optimized 4-fold higher in vivo solubility
140 Plasma Concentration (ng/mL)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 36 of 38
120 100 80 60 40 20 0 0
4
8
12 Time (hr)
626 627
ACS Paragon Plus Environment
16
20
24
Page 37 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
628
Molecular Pharmaceutics
Figure 6 (A)
629
(B)
630 631 632
ACS Paragon Plus Environment
Molecular Pharmaceutics
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
633
Figure 6 (C)
634
(D)
635 636 637 638 639
ACS Paragon Plus Environment
Page 38 of 38
