Development of Clinical Dosage Forms for a Poorly
`Water Soluble Drug I: Application of Polyethylene
`Glycol–Polysorbate 80 Solid Dispersion Carrier System
`
`ROSE-MARIE DANNENFELSER, HANDAN HE, YATINDRA JOSHI, SIMON BATEMAN, ABU T.M. SERAJUDDIN
`
`Novartis Pharmaceuticals Corporation, 1 Health Plaza, East Hanover, New Jersey 07936
`
`Received 11 September 2003; revised 23 November 2003; accepted 23 November 2003
`
`ABSTRACT: Different formulation approaches were evaluated to ensure that the formu-
`lation of a poorly water soluble compound chosen during early development achieves
`optimum bioavailability. The insoluble compound has an aqueous solubility of 0.17 mg/mL
`at 25  18C, a relatively high permeability (Caco2 Papp ¼ 6.1 104 cm/min), and poor
`bioavailability in dogs (dry blend formulation). Based on the prediction by GastroPlusTM,
`the oral absorption of this compound is sensitive to its apparent solubility and particle
`size. The oral bioavailability of three different formulations was compared in a dog model:
`a cosolvent-surfactant solution, a solid dispersion in a mixture of polyethylene glycol 3350
`and polysorbate 80, and a dry blend of micronized drug with microcrystalline cellulose.
`In absence of a parenteral injection, the bioavailability of the solution was considered to
`be 100%, and the relative oral bioavailability of the three formulations was 100, 99.1, 9.8,
`respectively. Comparable bioavailability was obtained with the solid dispersion and
`the cosolvent-surfactant solution, both of which showed a 10-fold higher bioavailability
`than the dry blend. Thus, a 20 mg dose strength capsule containing the solid dispersion
`formulation was selected for clinical development. The selected solid dispersion
`system was physically and chemically stable for at least 16 months at 258C/60% RH.
`In conclusion, the bioavailability of a poorly water soluble drug was greatly enhanced
`using the solid dispersion formulation containing a water soluble polymer with a surface
`active agent. ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci
`93:1165–1175, 2004
`Keywords: poorly soluble; solid dispersion; PEG; bioavailability enhancement
`
`INTRODUCTION
`
`Due to the application of combinatorial chemistry
`and high-throughput screening during drug dis-
`covery in recent years, a majority of new drug
`candidates exhibit poor aqueous solubility, result-
`ing in the development of bioavailable dosage
`forms for such compounds to be very challenging
`for formulation scientists. A poorly water-soluble
`
`Correspondence to: Rose-Marie Dannenfelser (Telephone:
`862-778-2069; Fax: 973-781-8487;
`E-mail: rosemary.dannenfelser@pharma.novartis.com)
`
`Journal of Pharmaceutical Sciences, Vol. 93, 1165–1175 (2004)
`ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association
`
`compound has classically been defined as one
`dissolving in less than 1 part per 10000 parts of
`water,1 which, in other words, is less than 100 mg/
`mL. A concentration of 100 mg/mL, however, may
`not appear to be very low in comparison with the
`aqueous solubility of <1 mg/mL (1 part per million)
`for compounds that in recent years are commonly
`emerging from drug discovery pipelines. A poorly
`water-soluble drug, more recently, has been
`defined in general terms to require more time to
`dissolve in the gastrointestinal fluid than it takes
`to be absorbed in the gastrointestinal tract.2 Thus,
`a greater understanding of dissolution and ab-
`sorption behaviors of drugs with low aqueous
`solubility is required to successfully formulate
`them into bioavailable drug products.
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
`
`1165
`
`Apotex v. Cellgene - IPR2023-00512
`Petitioner Apotex Exhibit 1040-0001
`
`

`

`1166
`
`DANNENFELSER ET AL.
`
`In efforts to enhance a drug’s dissolution rate
`and, in some cases, apparent aqueous solubility
`under gastrointestinal pH conditions, various
`techniques have been applied, including, but not
`limited to, particle size reduction. However, there
`is a practical limit with particle size reduction
`in how much can be achieved by conventional
`approaches.3 Therefore, formulation approaches
`are being explored to enhance bioavailability of
`poorly water-soluble drugs. One such formulation
`approach that has been shown to significantly
`enhance absorption of such drugs is the use of solid
`dispersions.3 – 18
`This article provides a case study leading to
`the development of a bioavailable dosage form for
`a poorly water-soluble drug, LAB687, for early
`phase clinical trials. The formulation approaches
`that have been investigated include a dry blend
`of micronized drug with excipients, a cosolvent-
`surfactant solution, and a solid dispersion system.
`Based on the relative oral bioavailability data in
`dog models, a solid dispersion formulation utiliz-
`ing a mixture of polyethylene glycol (PEG) 3350
`and polysorbate 80 as the drug carrier was chosen
`for Phase I clinical trials. Attempts were also
`made to correlate the drug’s pharmacokinetic data
`with the absorption predicted by GastroPlusTM
`(SimulationsPlus Inc., Lancaster, CA), a computer
`software that simulates and models gastroin-
`testinal absorption processes.
`
`MATERIALS AND METHODS
`
`Materials
`
`LAB687 (Fig. 1) is a neutral compound (MW 468.5)
`with a log P value of 4.7. It was obtained from the
`Chemical and Analytical Development depart-
`ment of Novartis Pharmaceuticals Corp. either in
`its polymorphic form A or as a mixture of poly-
`morphic forms A and C. The melting points for
`forms A and C are 131 and 1228C, respectively.
`The mixture of polymorphic forms A and C, which
`
`Figure 1. Structure of LAB687.
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
`
`has an aqueous solubility of 0.17 mg/mL at room
`temperature, was used in the present study. PEG
`3350 (Van Waters & Rogers, Glen Rock, NJ),
`polysorbate 80 (Ruger Chemical, Irvington, NJ),
`microcrystalline cellulose (Avicel PH 101, FMC,
`Philadelphia, PA), fumed silica (Cab-o-sil, Cabot,
`Boyertown, PA), and other tablet excipients were
`obtained by Novartis from commercial sources
`and released for human use. All other chemicals
`were used as received.
`
`HPLC Analysis
`
`An isocratic HPLC assay method using a reverse-
`phase column (Waters Symmetry C18, 7.5 cm 
`4.6 mm, 3.5 mm) at 308C and an UV detector set at
`a wavelength of 235 nm was used to analyze
`concentrations of the drug and its degradation
`products. The mobile phase consisted of a 50:50 v/
`v mixture of acetonitrile and water. At a flow rate
`of 1.5 mL/min, the retention time of LAB687 was
`5 min.
`
`Solubility Studies
`
`The solubility of LAB687 in various pharmaceu-
`tically relevant solvents were studied at 25  18C.
`Its solubility in water as a function of PEG 3350
`and polysorbate 80 concentrations at 37  18C was
`also studied. For those vehicles that were solid at
`room temperature, solubility was determined at
`60  58C, where the solvents were molten. Each
`solution was equilibrated for at least 24 h on a
`bottle rotator set at 40 RPM. Aliquots of solutions
`in organic solvents and those in aqueous media
`were centrifuged and then filtered through Milli-
`pore filters of 0.22 mm pore size before analyzing
`by HPLC.
`
`Solubility in Bile Salt–Lecithin Solutions
`
`Different amounts of lecithin were dissolved in
`40 mM sodium glycocholate solutions. Excess
`drug was equilibrated with the bile salt–lecithin
`solutions at 37  18C for at least 24 h prior to
`HPLC analysis.
`
`Dosage Form Development
`
`formulation approaches were
`Three different
`investigated to obtain the optimal formulations
`for Phase I clinical trials: a dry blend consisting of
`micronized drug, a solid dispersion, and an oral
`cosolvent-surfactant solution.
`
`Apotex v. Cellgene - IPR2023-00512
`Petitioner Apotex Exhibit 1040-0002
`
`

`

`DEVELOPMENT OF DOSAGE FORMS FOR A POORLY WATER SOLUBLE DRUG I
`
`1167
`
`Compatibility Screening
`
`A drug-excipient compatibility screening study
`was carried out to identify suitable excipients for
`the dry blend formulation. Mixtures of the drug
`with lactose, microcrystalline cellulose, stearic
`acid, magnesium stearate, and fumed silica in the
`presence of 20% added water were stressed at
`50  18C, as per the method described earlier.19
`Stability of LAB687 in different vehicles including
`solid dispersion carriers, such as PEG 400, PEG
`3350, propylene glycol, poloxamer 188, and poly-
`sorbate 80, was evaluated by storing the drug
`solutions at 50  58C for 3 weeks.
`
`Physical Stability
`
`In a preliminary study, it was established by
`spiking the solid dispersion matrices with crystal-
`line drug (1%) that the formation of a small
`number of drug crystals in the product during
`stability testing may not be detectable by powder
`X-ray diffraction and differential scanning calori-
`metry (DSC). However, an optical light micro-
`scope
`(Axioskop) with cross-polarized light
`connected to a hot stage (Mettler FP82) was used
`successfully to analyze the solid dispersions for
`the presence or absence of crystalline drug.
`Samples taken from the top, middle, and bottom
`of the solid dispersion plug were spread onto a
`glass slide and inserted into the hot stage.
`Observations were made initially, as is, at room
`temperature with cross-polarized light and as the
`samples were heated up to 708C at a heating rate
`of 58C/min on the hot stage. Crystalline material
`and all thermal events were noted.
`
`Prototype Formulations
`
`The dry blend formulation consisted of 50% micro-
`nized drug (mean particle size: 4.9 mm), 49.8%
`microcrystalline cellulose, and 0.2% fumed silica.
`Target fill weight for each of the size 0 hard
`gelatin capsules was 200 mg. The solid dispersion
`formulation consisted of 4% w/w LAB687 dis-
`persed in a 3:1 mixture of PEG 3350 and poly-
`sorbate 80. For this purpose, the drug was first
`dissolved in the molten carrier (PEG 3350/poly-
`sorbate 80) at 65  58C and a 500 mg aliquot of the
`hot molten solution was then manually filled with
`a positive displacement pipette into each size 0
`hard gelatin capsule. Each capsule thus contained
`20 mg of drug and 480 mg of carrier. The capsules
`were placed into high density polyethylene
`(HDPE) bottles and placed on stability.
`
`For bioavailability testing in animal models,
`50 mg of LAB687 was administered to each dog.
`Therefore, for ease of administration to dogs, sizes
`of powder-filled and solid dispersion-filled cap-
`sules were changed, although the compositions
`remained unchanged. Each size 000 capsule con-
`tained either 100 mg of powder blend or 1250 mg of
`solid dispersion.
`For the cosolvent-surfactant solution, 20 mg of
`drug was dissolved per milliliter of the cosolvent
`system consisting of 10% propylene glycol, 45%
`Cremophor RH40, 35% corn oil glycerides, and
`10% ethanol w/w/w/v. Each size 000 hard gelatin
`capsule was filled with 1.25 mL of the resulting
`solution and sealed with hot gelatin to eliminate
`leakage from the capsule.
`
`In Vitro Dissolution Testing
`
`Dissolution profiles of the capsule formulations
`containing micronized drug, solid dispersion, and
`cosolvent-surfactant solution were determined in
`0.01 N hydrochloric acid with 1% sodium lauryl
`sulfate (SLS) as the surfactant (pH 2), according
`to the USP apparatus I, basket method (100 RPM
`at 37  0.58C). Aliquots of dissolution medium
`collected at different time intervals were filtered
`through 0.45 mm filters and analyzed by HPLC.
`Additional dissolution studies with solid disper-
`sion capsules were also investigated using only
`water as the dissolution media with paddles at
`75 RPM, and in this case, aliquots were analyzed
`before and after filtration. Selected unfiltered
`aliquots were also analyzed by photon correla-
`tion spectroscopy using laser light scattering
`(Beckman Coulter N4 plus) to measure particle
`sizes of any phase-separated material in water.
`For this purpose, samples were loaded into 1 cm2
`cuvettes and placed in a thermostatic chamber.
`The sample viscosity and the water refractive
`index were factored in the particle size measure-
`ment by the instrument software. Light scattering
`was monitored at 908 angle and at a temperature
`of 258C.
`
`Absorption Prediction by GastroPlusTM
`
`LAB687 has a relatively high permeability (Caco2
`Papp ¼ 1.0  105 cm/s) and very low solubility
`(<1 mg/mL at pH 7.4). In Caco-2 cell incubations,
`the permeability of LAB687 from apical-to-baso-
`lateral cell surface was similar to that from
`basolateral-to-apical, suggesting that carrier-
`mediated efflux is not involved. Absorption of
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
`
`Apotex v. Cellgene - IPR2023-00512
`Petitioner Apotex Exhibit 1040-0003
`
`

`

`1168
`
`DANNENFELSER ET AL.
`
`LAB687 in the capsule formulations containing
`micronized drug, solid dispersion, and cosolvent-
`surfactant solution in dogs was predicted by
`GastroPlusTM (SimulationsPlus Inc.), a computer
`software that simulates and models the gastro-
`intestinal absorption processes based on the
`Advanced Compartmental Absorption and Transit
`(ACAT) model. Sensitivity analysis was also
`conducted to evaluate the relationships among
`absorption, solubility, particle size and dissolu-
`tion profiles.
`
`Bioavailability Study
`
`LAB687 was administered orally to three beagle
`dogs at a dose of 50 mg with one size 000 capsule
`containing micronized drug with excipients,
`the PEG 3350/polysorbate 80 solid dispersion, or
`two size 000 capsules of 20 mg/mL cosolvent-
`surfactant solution. A washout period of 5 days
`was used between dosing. The dogs were fasted
`overnight before dosing and allowed free access to
`water throughout the day. Plasma samples were
`analyzed for drug assay by LC/MS/MS method.
`The relative oral bioavailability (Frel) of each
`formulation was compared to that achieved from
`the reference, cosolvent-surfactant solution, and
`appropriate statistical analysis was conducted.
`
`RESULTS AND DISCUSSION
`
`Dosage Form Development
`
`LAB687 was found to be compatible with the
`evaluated tablet and solution excipients with no
`significant degradation products seen at 508C
`with 20% water for 1 week and 508C for 3 weeks,
`respectively. LAB687 was practically insoluble
`in water (solubility 0.17 mg/mL at 25  18C) and
`highly soluble in many pharmaceutically accep-
`table organic solvents. The solubility of the drug
`in PEG 400 was greater than 60 mg/mL at 258C.
`In a 3:1 mixture of PEG 3350 and polysorbate 80
`the solubility at 608C was greater than 100 mg/g
`(>10% w/w). To minimize any potential for the
`crystallization of drug at room temperature, a
`4% w/w solid dispersion was selected for further
`evaluation. Advantages of such a PEG-polysor-
`bate based solid dispersion carrier was previously
`reported in the literature.3,18,20,21
`
`In Vitro Dissolution
`
`Dissolution studies were conducted for purposes
`of quality control as well as for understanding
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
`
`the mechanism of release of drug from the dosage
`forms. Since the drug was insoluble in water
`throughout the gastrointestinal pH range, the
`dissolution profiles of prototype capsules contain-
`ing the micronized drug and the solid dispersion
`were determined in an acidic medium (0.01 N
`HCl) containing 1% SLS at 378C, and the results
`are shown in Figure 2a. Complete release of
`drug was observed in less than 45 min for the
`solid dispersion while only 80% in 45 min was
`released with the dry blend capsules. With the
`presence of a high concentration of SLS in the
`medium which is required for sink conditions,
`the dissolution method was developed for the
`batch-to-batch quality control of formulations and
`does not necessarily reflect in vivo performance
`of the dosage forms.
`For a better understanding of the release
`behavior of the drug in gastrointestinal fluid in
`absence of a high concentration of surfactant, the
`dissolution of the above-mentioned solid disper-
`sion formulation containing 25% w/w polysorbate
`80 in the matrix was studied in 250 mL of pure
`water at 378C, where the aliquots were analyzed
`before and after filtration through 0.22 mm filters,
`and the results are shown in Figure 2b. The
`medium (water) turned milky white in color,
`indicating that the solid dispersion mixed well
`with water even in the absence of added surfactant
`in the dissolution medium. Although PEG 3350
`and polysorbate 80 may not be miscible in all
`proportions,22 the amount of water used as the
`dissolution medium was sufficient to dissolve all
`of the PEG 3350 and polysorbate 80 in the formu-
`lation to create a clear solution.23 The milky color
`in the dissolution medium can thus be attributed
`to the active material being dispersed into very
`fine particles. The filtration of the dissolution
`medium revealed that approximately 20% of the
`drug passed through the 0.22 mm filter, and the
`concentration of drug in the filtrate remained
`practically the same whether a 0.22 or a 0.45 mm
`filter was used. As per the photon correlation
`spectroscopy using laser light scattering, the
`average sizes of the particulates in the unfiltered
`medium at 10, 30, and 60 min were 0.6  0.08,
`0.9  0.13, and 1.0  0.07 mm, respectively, while
`there were no measurable particles in the filtrate.
`It may be concluded from this experiment that
`under physiological conditions where approxi-
`mately 250 mL of gastrointestinal fluid might
`be present, about 20% of drug from the prototype
`solid dispersion capsule would dissolve in the
`medium and the excess would precipitate out as
`
`Apotex v. Cellgene - IPR2023-00512
`Petitioner Apotex Exhibit 1040-0004
`
`

`

`DEVELOPMENT OF DOSAGE FORMS FOR A POORLY WATER SOLUBLE DRUG I
`
`1169
`
`Figure 2.
`(a) Dissolution profiles for 50 mg/capsule dry blend and solid dispersion
`prototype formulations in 1000 mL of 0.01 N HCl/1% SLS using USP apparatus I, basket
`method at 100 RPMs and at 378C. (^: solid dispersion, &: dry blend); (b) dispersion study
`of the solid dispersion formulation in 250 mL water at 378C before and after filtration
`through 0.22 mm filter (^: 0.22 mm, &: nonfiltered); effect of polysorbate 80 concentration
`on the dispersibility of LAB687 in 1000 mL of water at 378C—percent recovery (c) after
`filtration through 0.45 mm filter and (d) before filtration. (^: 0% polysorbate 80, &: 5%,
`~: 10%, &: 15%, *: 20%, *: 25%).
`
`fine particles. The sizes of precipitated particles
`which were initially 0.6 mm (10-min point) in-
`creased with time (30 and 60 min points).
`To further elucidate the role of polysorbate 80
`in the solid dispersion carrier, the dissolution or
`dispersion of this formulation with different ratios
`of PEG 3350/polysorbate 80 (100:0 to 75:25) was
`studied in water (1000 mL) at 378C. Figure 2c
`shows the dissolution profiles (fraction passing
`through 0.45 mm filters) of LAB687 as a function of
`polysorbate 80 concentration in the matrix, while
`Figure 2d shows the dissolution/dispersion profiles
`in water when the aliquots were not filtered.
`The fraction passing through the filter, that is,
`the drug dissolved in the dissolution medium
`(water), increased gradually with an increase in
`polysorbate 80 concentration in the solid disper-
`sion matrix, and the total amount dispersed in
`water (unfiltered concentration) also increased
`with the increase in the surfactant concentration
`in the matrix. In a separate experiment, the solid
`dispersion formulation in neat PEG 3350 (no
`surfactant present in the matrix) was dispersed
`in 250 mL of water, and the particle size of the
`precipitated material was measured. Unlike the
`
`solid dispersion containing 25% w/w polysor-
`bate 80, the average size of the precipitated drug
`particles was 6–7 mm. Thus, the presence of sur-
`factant in the solid dispersion matrix is not only
`helpful in dissolving the drug in the dissolution
`medium, it also reduces the size of precipitated
`particles and thereby facilitates their redissolu-
`tion rate as a result of higher surface area.
`
`Stability
`
`The solid dispersion formulation (3:1 PEG 3350/
`polysorbate 80) showed excellent chemical and
`physical stability with minor changes in dissolu-
`tion (Table 1). In the solid dispersion formulation,
`the drug was either in a molecularly dispersed
`state or in a phase-separated amorphous state,
`since no crystals were detected after its prepara-
`tion. If it is in the amorphous state, there is a
`potential for crystallization of the drug substance
`to occur with time. Therefore, the solid dispersion
`formulation was monitored closely for any occur-
`rence of crystallization during the accelerated
`stability testing period. Although a formal stabi-
`lity study was conducted for 3 months only,
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
`
`Apotex v. Cellgene - IPR2023-00512
`Petitioner Apotex Exhibit 1040-0005
`
`

`

`1170
`
`DANNENFELSER ET AL.
`
`Table 1. Stability of the Solid Dispersion Capsules of LAB687 in a 3:1 Mixture of PEG 3350 and Polysorbate
`80 Stored in HDPE Bottles
`
`Time
`Point
`
`Initial
`1 M
`2 M
`3 M
`
`16 M
`
`Storage Condition
`
`Compound
`Potency (mg/cap)
`
`Total
`Impurity %
`
`Microscopic
`Analysis
`
`—
`408C/75% RH
`408C/75% RH
`258C/60% RH
`408C/75% RH
`258C/60% RH
`408C/75% RH
`
`20.1
`20.4
`20.6
`20.5
`20.5
`ND
`ND
`
`0.1
`0.1
`0.1
`0.1
`0.1
`ND
`ND
`
`No crystals
`No crystals
`No crystals
`No crystals
`No crystals
`No crystals
`Crystals
`
`% Dissolved
`
`15
`
`17.9
`ND
`ND
`ND
`
`37.6
`18.4
`
`30
`
`78.9
`ND
`ND
`ND
`
`100.2
`49.8
`
`45
`
`60
`
`103.2
`ND
`ND
`ND
`
`103.1
`74.2
`
`104.5
`103.1
`101.1
`99.3
`
`103.4
`84.8
`
`ND, not determined.
`
`subsequent visual examination of retained sam-
`ples did not show any crystallization of drug in the
`matrix even after 16 months of storage at 258C/
`60% RH. This indicates that the drug appears to
`have a low crystallization potential in the solid
`dispersion formulation. Only a small decrease in
`the dissolution rate of the capsules stored at 408C/
`75% RH was observed after 16 months (Table 1).
`It is a common phenomenon for hard gelatin
`capsules to show a decrease in dissolution rates
`when stored at 408C/75% RH.24 This phenomenon
`can be the main contributor to the decreased
`dissolution seen with the solid dispersion formu-
`lation, since only a few crystals were observed in
`the samples stored at 408C/75% RH for 16 months.
`
`GastroPlusTM Simulation
`The absorption in dogs predicted by GastroPlusTM
`for the 50 mg capsule formulation containing
`micronized drug utilizing an aqueous solubility of
`1 mg/mL and the drug particle size of 5 mm was low
`(8.6%) as shown in Figure 3 and the total fraction
`of absorption in the small intestine and colon
`was estimated at 1.5 and 7.1%, respectively. The
`absorption values were reflective of the drug
`transit time in the small intestine (1.5–1.8 h)
`and in the colon (10–12 h) in dogs, indicating
`that the drug was absorbed uniformly throughout
`the entire intestinal tract. As reported earlier, the
`drug solubility in water is 0.17 mg/mL, and,
`therefore, the use of a higher value of 1 mg/mL in
`the GastroPlusTM calculation was based on the
`consideration that the solubility of LAB687 in the
`gastrointestinal fluids would be higher in pre-
`sence of bile salts, lecithin, lipolytic products,
`etc.25 For example, the solubility of LAB687 was
`shown to be five times greater in 40 mM sodium
`glycocholate (0.8 mg/mL) than in water and almost
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
`
`10 times greater in the presence of 40 mM sodium
`glycocholate with 4 mM lecithin (1.9 mg/mL).
`The selection of the drug solubility value for the
`simulation could, however, be somewhat arbi-
`trary in nature since concentrations of such phy-
`siological
`solubilizing agents may vary in
`different regions of the gastrointestinal tract
`and from person to person. As shown later in the
`paper, a sensitivity analysis for drug solubility
`might present a better picture of the overall effect.
`The extent of absorption in the GastroPlusTM
`simulation was low primarily due to the poor
`solubility of the drug. A sensitivity analysis using
`the software showed that the oral absorption of
`LAB687 is very sensitive to solubility changes, as
`shown in Figure 4a. Though the actual solubility
`of the drug in the intestinal tract containing the
`mixture of the vehicle and intestinal fluid is
`unknown, the simulation suggested that different
`vehicles might have a significant impact on gastro-
`intestinal absorption of the drug. For example, if
`the solubility changes from 1 to 10 or 100 mg/mL,
`the absorption could increase from 8.6 to 50 or 95%,
`respectively. Based on the in vitro dissolution test,
`
`Figure 3. Predicted oral absorption of 50 mg LAB687
`in a dry blend with micronized drug by GastroPlusTM in
`dogs. Parameters used in the prediction include particle
`size of 5 mm and solubility value of 1 mg/mL.
`
`Apotex v. Cellgene - IPR2023-00512
`Petitioner Apotex Exhibit 1040-0006
`
`

`

`DEVELOPMENT OF DOSAGE FORMS FOR A POORLY WATER SOLUBLE DRUG I
`
`1171
`
`Figure 4. Simulated correlation for LAB687 in dogs between (a) absorption and
`solubility using a solubility range of 1–100 mg/mL, particle size of 4.9 mm, dose of 50 mg/
`dog, high permeability; (b) absorption and particle size using a particle size range of
`1–100 mm, solubility of 1 mg/mL, dose of 50 mg/dog, high permeability; (c) absorption and
`dose using a dose range of 5–500 mg/dog, solubility of 1 mg/mL, particle size of 4.9 mm, high
`permeability; (d) absorption and permeability using a dose of 50 mg/dog, particle size of
`4.9 mm, solubility of 1 mg/mL.
`
`it was found that the kinetic solubility and possibly
`even the equilibrium solubility of the drug in
`simulated intestinal fluid from a solid dispersion
`formulation could be increased as compared to that
`of the dry blend with micronized drug. Therefore,
`it is likely that the high bioavailability of the drug
`could be increased as compared to a dry blend by
`such formulation approaches as solid dispersion
`and cosolvent-surfactant solubilization. Similarly,
`the absorption is also sensitive to particle size
`changes, Figure 4b. On the other hand, due to the
`extremely low solubility of the drug, the absorption
`is not very sensitive to dose changes with a dose
`range between 5 and 500 mg/dog, Figure 4c. As
`expected, the oral absorption is relatively in-
`sensitive to permeability since the drug is highly
`permeable, Figure 4d.
`
`Dog Bioavailability Results
`
`The individual and mean serum concentrations
`and pharmacokinetic parameters of LAB687 in
`dogs are listed in Tables 2–5.
`The administration of the cosolvent-surfactant
`solution to the dogs resulted in a low inter-animal
`variability in concentrations of LAB687 as shown
`
`in Table 2. The absorption was rapid with a tmax of
`1–2 h. The Cmax and AUC0– 48h values ranged
`between 989 and 1320 ng/mL and between 6320
`and 7540 ng h/mL, respectively. When the cosol-
`vent-surfactant solution was diluted with water,
`the particle size measured by laser light scattering
`was 125 mm, possibly representing a micellar
`system or microemulsion. The AUC0 – 48h (6960 ng
`h/mL) of this formulation was, therefore, used as
`the reference for calculating the relative bioavail-
`ability (Frel) of the other two formulations since
`the drug could not be given for an intravenous
`dose due to lack of a biocompatible intravenous
`formulation.
`The administration of solid dispersion to dogs
`also resulted in a low inter-animal variability
`in the concentrations of LAB687 (Table 3) except
`that the concentration at 12 h in Dog 2 was higher
`than that in the other dogs. The drug was rapidly
`absorbed with the time to peak concentration at
`1–2 h. The Cmax and AUC0– 48h values ranged
`between 770 and 853 ng/mL and between 5890 and
`8900 ng h/mL, respectively. The mean AUC0– 48h of
`6900 ng h/mL showed an equal relative bioavail-
`ability of LAB687 (99.1%) as compared to the
`cosolvent-surfactant solution.
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
`
`Apotex v. Cellgene - IPR2023-00512
`Petitioner Apotex Exhibit 1040-0007
`
`

`

`1172
`
`DANNENFELSER ET AL.
`
`Table 2. Serum Concentrations and Relevant Pharmacokinetic Parameters
`Following Single Oral Doses of 50 mg/dog LAB687 in Cosolvent-Surfactant Solution
`
`Concentration (ng/mL)
`
`Time (h)
`
`Dog 1
`
`Dog 2
`
`Dog 3
`
`Mean
`
`0
`0.5
`1
`2
`4
`8
`12
`24
`48
`Tmax (h)
`Cmax (ng/mL)
`AUC0– 48h (ng : h/mL)
`
`BLQa
`279
`989
`820
`240
`169
`137
`112
`32.4
`1
`989
`7010
`
`BLQ
`615
`1320
`574
`296
`231
`167
`42.0
`21.8
`1
`1320
`6320
`
`BLQ
`35.3
`602
`1170
`383
`249
`189
`78.0
`21.3
`2
`1170
`7540
`
`BLQ
`310
`970
`855
`306
`216
`164
`77.3
`25.2
`1–2b
`1160
`6960
`
`aBLQ indicates concentration below limit of quantification (<21.2 ng/mL).
`bA range was given.
`cNot applicable.
`
`SD
`
`BLQ
`291
`359
`300
`72.1
`42.0
`26.1
`35.0
`6.27
`n/ac
`166
`610
`
`In contrast to the formulations described above,
`the administration of dry blend with micronized
`drug to dogs resulted in a high inter-animal
`variability in concentrations of LAB687 as shown
`in Table 4. It was found that the concentrations
`of the parent compound in one dog (Dog 1) were
`significantly higher than those in the other dogs.
`The absorption showed a rapid onset after the
`administration with a tmax of 1 h. However, the
`Cmax and AUC0 – 48h were low with values ranging
`between 43.7 and 245 ng/mL and between 184 and
`1650 ng h/mL, respectively. The mean Frel value
`
`of the formulation showed a low relative bio-
`availability of LAB687 (9.8%) as compared to the
`solid dispersion or cosolvent-surfactant solution
`(Table 5).
`Although the absolute bioavailability of LAB687
`in these formulations could not be determined, the
`predicted oral bioavailability from the dry blend
`with micronized drug (8.6%) correlated well with
`the in vivo relative oral bioavailability (9.8%),
`suggesting that the drug in the cosolvent-surfac-
`tant solution or solid dispersion possibly provided
`almost complete bioavailability, which was tenfold
`
`Table 3. Serum Concentrations and Relevant Pharmacokinetic Parameters
`Following Single Oral Doses of 50 mg/dog LAB687 in Solid Dispersion Matrix
`
`Concentration (ng/mL)
`
`Time (h)
`
`Dog 1
`
`Dog 2
`
`Dog 3
`
`Mean
`
`0
`0.5
`1
`2
`4
`8
`12
`24
`48
`Tmax (h)
`Cmax (ng/mL)
`AUC0– 48h (ng : h/mL)
`
`BLQa
`BLQ
`430
`787
`277
`196
`133
`74.2
`33.0
`2
`787
`5910
`
`BLQ
`179
`527
`853
`292
`167
`523
`62.6
`23.6
`2
`853
`8900
`
`BLQ
`61.8
`770
`718
`365
`242
`221
`62.0
`BLQ
`1
`770
`5890d
`
`BLQ
`80.3
`576
`786
`311
`202
`292
`66.3
`18.9
`1–2b
`803
`6900
`
`SD
`
`BLQ
`90.9
`175
`67.5
`47.1
`37.8
`205
`6.88
`17.0
`n/ac
`43.8
`1730
`
`aBLQ indicates concentration below limit of quantification (<21.2 ng/mL).
`bA range was given.
`cNot applicable.
`dArea under serum concentration–time curve from 0 to last measurable time point.
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 5, MAY 2004
`
`Apotex v. Cellgene - IPR2023-00512
`Petitioner Apotex Exhibit 1040-0008
`
`

`

`DEVELOPMENT OF DOSAGE FORMS FOR A POORLY WATER SOLUBLE DRUG I
`
`1173
`
`Table 4. Serum Concentrations and Relevant Pharmacokinetic Parameters
`Following Single Oral Doses of 50 mg/dog LAB687 in Dry Blend Containing
`Micronized Drug
`
`Concentration (ng/mL)
`
`Time (h)
`
`Dog 1
`
`Dog 2
`
`Dog 3
`
`Mean
`
`0
`0.5
`1
`2
`4
`8
`12
`24
`48
`Tmax (h)
`Cmax (ng/mL)
`AUC0 –48h (ng : h/mL)
`
`BLQa
`BLQ
`245
`146
`127
`48.3
`45.7
`51.9
`BLQ
`1
`245
`1650
`
`BLQ
`BLQ
`94.8
`41.4
`BLQ
`BLQ
`25.6
`BLQ
`BLQ
`1
`94.8
`184
`
`BLQ
`21.5
`43.7
`21.3
`27.4
`BLQ
`24.3
`BLQ
`BLQ
`1
`43.7
`205c
`
`BLQ
`7.20
`128
`70.0
`51.0
`16.0
`32.0
`17.3
`BLQ
`1
`128
`681
`
`SD
`
`BLQ
`12.4
`105
`67.0
`66.8
`27.9
`12.0
`30.0
`BLQ
`n/ab
`105
`842
`
`aBLQ indicates concentration below limit of quantification (<21.2 ng/mL).
`bA range was given.
`cArea under the serum concentration–time curve from 0 to last measurable time point.
`
`higher than that of the dry blend with micronized
`drug. Furthermore, the solid dispersion or cosol-
`vent-surfactant solution showed much less varia-
`bility in terms of the pharmacokinetic parameters.
`It appeared that the rate, extent, and gastrointest-
`inal location of oral absorption of the drug could be
`predicted well by GastroPlusTM a physiological
`based simulation software. The simulation pro-
`gram allowed us to incorporate the experimental
`data of the drug into the prediction (e.g., Caco-2
`permeability, solubility) along with the built-in
`physiological values (such as pH and transit time
`of each intestinal segment in dogs) to increase
`the reliability of the absorption prediction. Other
`absorption related factors including the time for
`the rate of precipitation, water volume after the
`oral dose in capsule, particle size were also
`considered in the prediction. The good Gastro-
`PlusTM prediction for LAB687 could also be factors
`of its minimal first pass metabolism and that it
`is not involved in any active transporters. Most
`importantly, the software provided a sensitivity
`analysis, that allowed users to understand the
`
`correlations between absorption and associated
`parameters, such as solubility, dose, particle size,
`permeability, and pharmacokinetic parameters.
`Thus, the software indeed provided an insight
`into the formulation development processes and
`allowed foresight to potential issues prior to for-
`mulation investigation.
`It should also be mentioned that according
`to Chiou and Buehler,26 drugs are often absorbed
`more completely in dogs than in humans, and,
`therefore, the oral absorption in dogs might not
`be a true reflection of hum