ELSEVIER (cid:9)
`
`International Journal of Pharmaceutics 157 (1997) 43-52
`
`international
`journal of
`pharmaceutics
`
`Design of sustained-release matrix systems for a highly
`water-soluble compound, ABT-089
`
`Yihong Qiu *, Kolette Flood, Kennan Marsh, Sherry Carroll, Jay Trivedi,
`Stephen P. Arneric, Steven L. Krill
`
`Pharmaceutical Products Division, Abbott Laboratories, 100 Abbott Park Rd., Abbott Park, IL 60064-3500, USA
`
`Received 28 April 1997; received in revised form 27 June 1997; accepted 2 July 1997
`
`Abstract
`
`ABT-089 is a potent cholinergic channel modulator under investigation for treatment of cognitive disorders. It is
`a highly water soluble compound with a short elimination half-life of 1.7 h in dogs. Hydrophilic and hydrophobic
`matrix systems were designed to investigate the feasibility of prolonged oral delivery of ABT-089 and to explore the
`preliminary in vitro and in vivo correlations. The sustained-release single and layered matrix tablets were prepared by
`compression. In vitro release testing using a USP apparatus II was performed for formulation screening. The release
`rates were modulated by varying concentrations of different types of rate controlling materials and by restricting
`surface area available for drug release. The transport mechanism of the compound from different types of systems
`typically followed Fickian diffusion. Based on the in vitro release characteristics, two types of prototype matrix
`systems were evaluated in beagle dogs. Both formulations provided prolonged plasma levels of ABT-089 above the
`minimum effective concentration for over 22 h with reduced fluctuation of plasma levels. In vivo drug release from
`the tablet matrix estimated by deconvolution correlated well with drug release in vitro. In conclusion, prolonged oral
`delivery of highly soluble ABT-089 was achieved using diffusion controlled matrix systems. The hydrophobic matrix
`was found to be more effective than hydrophilic matrix in extending the release of the compound. Linear relationships
`between in vitro and in vivo drug release indicated by the initial results for both types of systems can provide useful
`information for further formulation development. © 1997 Elsevier Science B.V.
`
`Keywords: Cholinergic channel modulator; Sustained-release; Matrix system; In vitro release; Deconvolution; In
`vitro/in vivo correlation
`
`* Corresponding author. Abbott Laboratories, Formulation Development Center, D-4P7/R1B, 1401 N. Sheridan Rd., North
`Chicago, IL 60064, USA. Tel: + 1 847 9385220; fax: + 1 847 9351997; e-mail: yihong.qiu@abbott.com
`
`0378-5173/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved.
`PH S0378-5173(97)00215-9
`
`MYLAN Ex 1045, Page 1
`
`

`

`44 (cid:9)
`
`Y. Qiu et al. /International Journal of Pharmaceutics 157 (1997) 43-52
`
`1. Introduction
`
`[2-methy1-3-(2-(S)-pyrrolidinyl-
`ABT-089, (cid:9)
`methoxy)pyridine], is a potent and selective
`cholinergic channel modulator which is under in-
`vestigation for treatment of cognitive disorders
`(Williams and Arneric, 1996) (Fig. 1). Unlike
`(-)-nicotine, ABT-089 is substantially orally avail-
`able and has a reduced activation of human gan-
`glionic nicotinic receptors. Hence, it has a reduced
`propensity to elicit cardiac arrhythmias at high
`doses (Arneric et al., 1996). The compound is
`crystalline and thermally stable both in solution
`and in the solid state. ABT-089 is very soluble in
`aqueous media at pH values between 1-14 ( > 6
`g/ml). However, it is eliminated rapidly in vivo
`with a t1/2 of 1.7 h in dogs (Arneric et al., submit-
`ted). Multiple dosing is necessary to maintain
`plasma concentrations above minimum effective
`concentration (MEC) of 2 ng/ml (Arneric et al.,
`1996). In addition, preliminary studies have
`shown that prolonged exposure to ABT-089 with
`a reduced peak-to-trough ratio is therapeutically
`beneficial in selected animal models. Therefore, a
`sustained-release formulation which could be
`given once daily should be advantageous.
`In the present study, various matrix systems of
`ABT-089 were designed and tested for sustained
`release of ABT-089. The objectives of the study
`were (1) to investigate the feasibility of prolonged
`oral delivery of ABT-089, (2) to evaluate the
`performance of hydrophilic and hydrophobic ma-
`trix systems in retarding the release of this highly
`soluble compound, (3) to explore the preliminary
`relationship between in vitro and in vivo drug
`release from the matrix systems to facilitate fur-
`ther formulation development.
`
`2. Experimental section
`
`2.1. Materials and equipment
`
`The following materials were used in the study:
`ABT-089 • 2 HCl (Pharmaceutical Products Divi-
`sion, Abbott Laboratories) (Lin et al., 1997); Hy-
`droxypropyl methylcellulose, Methocel KlOOM
`(Dow Chemical Co.); Poly(ethyleneoxide),
`
`Polyox® coagulant (Union Carbide Co.); Car-
`nauba wax (J.W. Hanson Co., Inc.); Partially
`hydrogenated cottonseed oil, Stereotex K (Abitec
`Co.). All other chemicals and reagents were either
`AR or HPLC grades and used as received. A
`Vanderkamp® 600 dissolution tester and a
`HP8452A UV—VIS diode array spectrophotome-
`ter were used to determine the in vitro drug
`release. An HPLC system consisted of an Applied
`Biosystems 400 isocratic pump, an ABI 491 high-
`pressure dynamic mixer, a Hitachi 655A-40 au-
`tosampler and a Shimadzu RE-551 fluorescence
`detector with a Beckman PeakPro data collection
`system.
`
`2.2. Formulations
`
`2.2.1. Hydrophilic matrix system
`High viscosity grade hydroxypropyl methylcel-
`lulose (HPMC) and poly(ethyleneoxide) were used
`to prepare single and/or layered matrix tablets by
`direct compression using a Carver hydraulic press.
`Their compositions are described in Table 1.
`
`2.2.1.1. Single tablet. ABT-089 .2 HC1 was dry
`mixed with the polymer and other excipients. The
`blend was directly compressed into a 300-mg
`tablet using a concave punch at 4000 lb with a
`dwell time of 5 s.
`
`2.2.1.2. Layered tablet. Two types of layered
`tablets were designed, i.e. the hydrophilic matrix
`containing the drug is coated with hydrophilic
`barriers on both faces of the tablet by compres-
`sion (HHH) and the hydrophilic matrix contain-
`ing the drug is coated with hydrophobic barriers
`on both faces of the tablet by compression
`(WHW). Ingredients of the middle and barrier
`
`C11H16N20, MW = 192.28
`
`Fig. 1. Chemical structure of ABT-089.
`
`MYLAN Ex 1045, Page 2
`
`

`

`Y. Qiu et al. /International Journal of Pharmaceutics 157 (1997)
`
`43-52
`
`45
`
`Table 1
`Hydrophilic matrix system components
`
`Dosage form
`
`Single tablet (cid:9)
`
`Layered tablet
`
`Label
`
`H1
`
`H2 (cid:9) HHH
`
`WHW
`
`Top layera Middle layer
`
`Bottom layera Top layerb Middle layer
`
`Bottom layerb
`
`ABT-
`089.2 HC1
`(%)
`Methocel®
`K100M (%)
`Polyox® coagu-
`lant (%)
`Carnauba wax
`(%)
`Lactose anhy-
`drous (%)
`
`10
`
`26
`
`10
`
`51.8
`
`— (cid:9)
`
`54.5
`
`35.0
`
`54.5
`
`25
`
`51.8
`
`35.0
`
`10
`
`90
`
`10
`
`90
`
`64
`
`65 (cid:9)
`
`45.5
`
`13.2
`
`45.5
`
`13.2
`
`Total (mg)
`
`300
`
`300 (cid:9)
`
`110
`
`80
`
`110
`
`110
`
`80
`
`110
`
`aHydrophilic barrier layer.
`bHydrophobic barrier layer.
`
`layers were blended separately. Layered matrix
`tablets were prepared by compressing the barrier
`layer at 300 lb followed by middle layer at 300 lb
`and another barrier layer at 4500 lb using a
`flat-faced punch with a dwell time of 5 s.
`
`2.2.2. Hydrophobic matrix system
`Carnauba wax (W1 —W3) and partially hydro-
`genated cottonseed oil (W4—W7) were used as the
`rate controlling materials to prepare hydrophobic
`matrix tablets by compression using a Carver
`press. The formulations of the matrices are listed
`in Table 2. Drug and other excipients were
`blended and slowly added to molten wax at —
`95°C and mixed thoroughly. The mixture was
`allowed to congeal at room temperature while
`
`Table 2
`Hydrophobic matrix system components
`
`mixing. The congealed solids were milled and
`passed through a 30 mesh screen. The tablets were
`prepared with a concave punch by compressing at
`4500 lb with a dwell time of 5 s. The chemical
`stability of ABT-089 (m.p., 210°C) in the formula-
`tion was confirmed by potency assay.
`
`2.3. In vitro release
`
`The in vitro release tests were performed using
`the USP apparatus II (paddle method). The disso-
`lution medium was 900 ml of distilled water main-
`tained at 37 ± 0.5°C. The paddle rotation speed
`was kept at 100 rev./min. Water was used as the
`initial dissolution medium because of extremely
`high solubility of ABT-089 at different pH. In all
`
`Formulation
`
`ABT-089 2 HC1 (%)
`Carnauba wax (%)
`Sterotex K wax (%)
`Lactose anhydrous (%)
`
`Total (mg)
`
`W1
`
`13.8
`60
`
`26.2
`
`300
`
`W2
`
`13.8
`55
`
`31.2
`
`300
`
`W3
`
`13.8
`50
`
`36.2
`
`300
`
`W4
`
`13.8
`
`55
`31.2
`
`300
`
`W5
`
`13.8
`
`51
`35.2
`
`300
`
`W6
`
`13.8
`
`48
`38.2
`
`300
`
`W7
`
`13.8
`
`40
`46.2
`
`300
`
`MYLAN Ex 1045, Page 3
`
`(cid:9)
`(cid:9)
`

`

`46 (cid:9)
`
`Y. Qiu et al. /International Journal of Pharmaceutics 157 (1997) 43-52
`
`Fractional Release
`
`Fig. 2. In vitro release profiles of different types of hydrophilic matrices of ABT-089 containing [ x ] 26% Methocel K100M (H1);
`[q
`] 25% Polyox coagulant (H2); [+] 35% Methocel K100M with hydrophilic barrier layers (HHH) and [0] 35% Methocel K100M
`with hydrophobic barrier layers (WHW) compared with a hydrophobic matrix containing [•] 55% Carnauba wax (W2) (24 h data
`not shown).
`
`experiments, 3.0 ml dissolution samples were with-
`drawn at predetermined time intervals for up to 24
`h, and replaced with equal volumes of the fresh
`medium to maintain the total volume constant.
`Samples were filtered through a filter (4.5 pm) and
`assayed by UV spectrophotometry at 276 nm.
`
`2.4. In vivo studies
`
`Based on the in vitro release, a layered hy-
`drophilic matrix (HHH) and a simple hydropho-
`bic matrix (W2) with significantly differing in
`vitro release rates were assessed in a series of
`studies in a group of six beagle dogs. Animals
`were handled according to protocols approved by
`Abbott's Institutional Animal Care and Use
`Committee. Each study was carried out at least 1
`week apart. An oral aqueous solution of ABT-089
`and an immediate-release capsule were used as
`references.
`The preliminary evaluation of the matrix for-
`mulations HHH was carried out in a sub-group of
`three animals. The animals were fasted overnight
`prior to dosing but were permitted water ad libi-
`
`tum. Each animal received a single tablet contain-
`ing 30 mg of ABT-089 followed by — 10 ml of
`water. Under this fasting dosing regimen, food
`was returned to each animal 12 h after dosing.
`A second study evaluated the matrix formula-
`tion W2 with the release rate slower than matrix
`HHH. The formulation was orally administered
`to a group of six fasting animals using the same
`protocol as described above.
`Sequential blood samples were obtained from
`each animal prior to dosing and at selected time
`points post dosing interval in each of the studies
`outlined above. The blood sampling schemes were
`0, 0.25, 0.5, 0.75, 1.0, 1.5, 2.0, 3.0, 4.0, 6.0, 9.0,
`12.0 h for the immediate-release formulations and
`0, 0.25, 0.5, 1.0, 1.5, 2.0, 4.0, 6.0, 9.0, 12.0, 15.0
`and 24.0 h for the sustained-release tablets, re-
`spectively. Plasma was separated by centrifuga-
`tion (2500 rev./min x 10 min, 4°C) and frozen
`( — 30°C) until analyzed. The plasma concentra-
`tions of ABT-089 were determined by a validated
`method of reverse phase HPLC following pre-
`column fluorescent derivatizati on (Hui and
`Marsh, 1995).
`
`MYLAN Ex 1045, Page 4
`
`

`

`Y. Qiu et al. /International Journal of Pharmaceutics 157 (1997) 43-52 (cid:9)
`
`47
`
`Table 3
`Results of linear regression of fractional release (F) vs square-root-of-time (t'/2) for hydrophilic and hydrophobic sustained-release
`matrices of ABT-089
`
`Matrix system
`
`Release rate constant (h -1/2)
`
`Intercept
`
`Coefficient of Determination (R2)
`
`H1
`HHH
`WHW
`
`W1
`W2
`W3
`W4
`W5
`W6
`W7
`
`0.449
`0.424
`0.355
`
`0.162
`0.188
`0.205
`0.115
`0.181
`0.201
`0.287
`
`0.105
`—0.082
`0.015
`
`0.043
`0.039
`0.013
`0.045
`0.071
`0.098
`0.071
`
`0.9983
`0.9752
`0.9947
`
`0.9952
`0.9985
`0.9967
`0.9928
`0.9940
`0.9952
`0.9906
`
`2.5. Data analysis
`
`3.1. In vitro release kinetics
`
`The area under the plasma concentration—time
`curve from time zero to the last sampling time
`point t (AUCt) was calculated by the trapezoidal
`rule. The AUC values were normalized on the
`basis of dog weights. Deconvolution was per-
`formed in order to evaluate the rate of drug
`release/absorption. Plasma concentration data fol-
`lowing oral administration of the solution were
`fitted to polyexponentials (PROC NLIN of SAS,
`version 6.09, of SAS Institute, Cary, NC and
`RSTRIP of Micromath®, Salt Lake City, UT) and
`used as the unit impulse response, C,5(t). Drug
`plasma data from tablet formulations C(t) were
`fitted to a smoothing cubic spline function and
`then deconvoluted with C,5(t) using program
`PCDCON (W.R. Gillespie, FDA) to estimate in
`vivo drug release from the matrix formulations.
`
`3. Results and discussion
`
`Physicochemical properties of a compound are
`important to drug absorption as well as the design
`of the delivery system. Due to extremely high
`aqueous solubility of ABT-089, hydrophobic and
`hydrophilic matrix systems with or without re-
`stricted release area were tested to control the
`drug release.
`
`Fig. 2 compares drug release profiles of differ-
`ent types of matrix formulations. Each curve typi-
`cally represents the mean of three replicates.
`Overall, low variability was observed in the re-
`lease profiles (CV < 5%).
`
`3.1.1. Hydrophilic matrix system
`The release of a dispersed drug from a non-
`crosslinked hydrophilic polymer matrix system
`can be related to time according to Eq. (1) (Ho-
`gan, 1989):
`
`F = kt" (cid:9)
`
`(1)
`
`where F is the fraction released at time t, k is a
`constant incorporating characteristics of the
`macromolecular network system and the drug,
`and n is an exponent characteristic of the trans-
`port mechanism. Eq. (1) is a generalized semi-em-
`pirical equation that describes two apparently
`independent mechanisms of drug transport from a
`matrix system, i.e. a Fickian and a non-Fickian
`mechanism. ABT-089 was homogeneously dis-
`persed throughout the hydrophilic polymer ma-
`trix. Because of its high solubility, Fickian
`diffusion is expected to be a predominant release
`mechanism. Thus, an approximately linear rela-
`tionship between fractional release and the square
`root of time was obtained for this system (Table
`3). The drug release from the single layer matrices
`
`MYLAN Ex 1045, Page 5
`
`(cid:9)
`

`

`48 (cid:9)
`
`Y. Qiu et al. /International Journal of Pharmaceutics 157 (1997) 43-52
`
`2
`
`4
`Time (hr)
`
`8
`
`1.0 (cid:9)
`
`
`
`0.8 -
`
`0.6
`
`0.4
`
`0.0
`
`0
`
`Fractional Release
`
`Fig. 3. Comparison of ABT-089 release profiles from hydrophobic matrices containing [A] 50% (W3); [0] 55% (W2); [n ] 60%
`Carnauba wax (W1), and [ x ] 40% (W7); [q
`] 48% (W6); [1] 51% (W5) and [0] 55% (W4) Sterotex K (24 h data not shown).
`
`was rapid. Approximately 90% of the drug was
`released in the first 3-4 h. Rapid release was
`likely due to the high solubility as well as high
`diffusivity of ABT-089 since two matrices con-
`taining polymers of differing viscosities resulted in
`the same release rates at a relatively high polymer
`concentration.
`
`3.1.2. Three-layer hydrophilic matrix system
`Drug release rate is not only dependent on
`diffusion coefficient, diffusion path length and
`concentration gradient, but also on the surface
`area available for release. Hence, reduction of
`releasing area was used to modulate the dissolu-
`tion rate of ABT-089. A hydrophilic matrix con-
`taining the drug was sandwiched with soluble/
`swellable barriers on both bases of the tablet by
`compression. In this system, the release of drug
`from the middle layer was effectively slowed by
`delaying or preventing drug diffusion from the
`two base surfaces. The early release of the drug
`occurred from the greatly reduced surface along
`the peripheral wall. Complete release can be
`achieved as a result of polymer dissolution over a
`prolonged period of time (e.g. 24 h). Fig. 2 indi-
`
`cates that the release rate of ABT-089 from the
`HHH system was significantly retarded. A differ-
`ent hydrophilic layered matrix, WHW, was pre-
`pared by applying hydrophobic insoluble barriers
`on both bases of the tablet. It was observed that
`the drug release rate of the system WHW was
`similar to that of HHH system. The fact that both
`insoluble and soluble barriers are equally effective
`in decreasing the release rate suggests that the
`middle layer hydrates and erodes more rapidly
`than the soluble barrier layers. Complete release
`was also obtained as the middle layer was com-
`pletely hydrated and eroded by 24 h. The drug
`release from matrix WHW also followed a square-
`root of time relationship. Although decreased re-
`lease rate had been obtained with the layered
`matrices, the overall drug release was still rela-
`tively fast, possibly a result of both high solubility
`and diffusivity of ABT-089. To further extend the
`release of ABT-089 from the dosage form, hydro-
`phobic matrix systems were tested.
`
`3.1.3. Hydrophobic matrix system
`Hydrophobic inert matrix is one of the well
`developed matrix systems used for sustained drug
`delivery because of its effectiveness and low cost,
`
`MYLAN Ex 1045, Page 6
`
`

`

`Y. Qiu et al. /International Journal of Pharmaceutics 157 (1997) 43-52 (cid:9)
`
`49
`
`y = 0.422 - 0.004x RA2 = 0.986
`
`y = 0.728 - 0.011x RA2 = 0.985
`
`0.3
`
`0.2 -
`
`Release Rate (hr^-1/2)
`
`0.1
`30
`
`40
`
`5'0
`Wax %
`
`601
`
`70
`
`Fig. 4. Release rates versus percent wax of different hydrophobic matrices ([ x ]: Carnauba wax; [0]: Sterotex K).
`
`1000
`
`10
`
`Plasma Concentration (ng/ml)
`
`4 (cid:9)
`
`8 (cid:9)
`
`12
`Time (hr)
`
`16
`
`20
`
`24
`
`Fig. 5. Plasma concentration-time profiles of ABT-089 in dogs after a single oral dose of [A] an immediate-release (IR) capsule,
`[0] a layered hydrophilic matrix containing 35% Methocel K100M (HHH) and [•] a hydrophobic matrix containing 55% carnauba
`wax (W2) in the fasted state; [—] MEC (mean ± S.E., Dose = 3.0 mg/kg).
`
`especially for highly water-soluble compounds.
`There have been many studies concerning the
`matrix system and its variations with modified
`release kinetics (Higuchi, 1963; Desai et al., 1965;
`Foster and Parrott, 1990; Scott and Hollenbeck,
`
`1991; Lee, 1992; Otsuka and Matsuda, 1994,
`1995; Hildgen and McMullen, 1995). Compared
`with other delivery systems, this type of matrix is
`relatively insensitive to changes in release environ-
`ment because diffusion from the noneroding ma-
`
`MYLAN Ex 1045, Page 7
`
`

`

`50 (cid:9)
`
`Y. Qiu et al. /International Journal of Pharmaceutics 157 (1997) 43-52
`
`o (cid:9)
`•
`
`y = 0.000 + 1.088x RA2 = 0.997
`y = - 0.006 + 0.888x RA2 = 0.988
`
`1.0
`
`0.8 -
`
`0.6 -
`
`0.4 -
`
`0.2 -
`
`Fractional release in vivo
`
`0.0 (cid:9)
`0 0 (cid:9)
`
`
`0.2
`
`0.6
`0.4 (cid:9)
`Fractional release in vitro
`
`0 . 8
`
`1 0
`
`Fig. 6. Correlations between in vitro and in vivo release of ABT-089 from [0] a layered hydrophilic matrix containing 35% Methocel
`K100M (HHH) and [•1 a hydrophobic matrix containing 55% carnauba wax (W2).
`
`trix is the rate controlling factor (Desai et al.,
`1965). The release mechanism of a heterogeneous
`matrix is diffusion through aqueous channels
`formed by dissolution of the drug and soluble
`excipients in the formulation. Based on sink con-
`ditions and pseudo-steady-state approximation
`(drug loading per unit volume, A, » solubility,
`Cs), Higuchi (1963) developed an approximate
`analytical solution for a rigid planar matrix, i.e.
`D
` t (cid:9)
`
`(2)
`
`1 [
`Mt
`Moc, Moc,
`
`cCs(2A— ECs)
`
`where Mt and Al„, are the amounts released per
`unit area at t and infinity, respectively. D is the
`diffusion coefficient of the drug in the medium, c
`is porosity and r is tortuosity of the matrix.
`In the present study, the use of a hydrophobic
`carnauba wax matrix significantly decreased re-
`lease rates of ABT-089 as demonstrated in Fig. 3.
`The release mechanism from the non-swellable
`matrix system followed typical Fickian diffusion
`that can be described by Eq. (2) (Table 3).
`Carnauba wax was found to be effective in
`controlling the release of ABT-089. However, the
`formulation suffered from rapid hardening and
`
`sticking which caused difficulty in cleaning equip-
`ment. In order to overcome this problem, par-
`tially hydrogenated cottonseed oil (Sterotex K),
`with a melting point similar to carnauba wax, was
`evaluated as a hydrophobic rate controlling mate-
`rial. Drug release profiles comparing hydrophobic
`matrix formulations containing different percent-
`ages of carnauba wax (W1 —W3) and Sterotex K
`(W4—W7) are shown in Fig. 3. The release rates
`of Formulations W1 —W7 were determined by
`fitting the in vitro release data to Eq. (2) (Table
`3). It was found that approximately linear rela-
`tionships between drug release rate and percent
`wax in the formulations were obtained for car-
`nauba wax as well as Sterotex K formulations
`(Fig. 4). As can be seen, the formulations contain-
`ing carnauba wax resulted in a comparatively
`higher release rate than those containing Sterotex
`K at the concentrations above approximately
`46%. The rate of drug release also changed more
`rapidly with wax concentration for formulations
`containing Sterotex K. This observation may be
`attributed, at least in part, to the difference in
`hydrophobicity between the two wax materials.
`More hydrophobic trigylerides (Sterotex K) might
`
`MYLAN Ex 1045, Page 8
`
`

`

`Y. Qiu et al. /International Journal of Pharmaceutics 157 (1997) 43-52 (cid:9)
`
`51
`
`be more effective in preventing the penetration of
`water and/or interaction with water than esters of
`hydroxylated unsaturated fatty acid (Carnauba
`wax).
`
`3.2. In vivo absorption
`
`Fig. 5 shows the plasma concentration—time
`profiles of ABT-089 following single oral adminis-
`tration of the two matrix formulations in dogs as
`compared with an immediate-release capsule for-
`mulation. Prolonged absorption was achieved
`with both sustained-release formulations, provid-
`ing plasma levels of ABT-089 above the minimum
`effective concentration (Arneric et al., 1996) of 2
`ng/ml for over 22 h. Qualitative comparison of
`the two formulations relative to the reference
`suggested that the sustained effects were in the
`same rank order as the in vitro drug release.
`Preformulation and animal studies had indicated
`that ABT-089 is stable at intestinal pH and effi-
`ciently absorbed throughout the whole intestinal
`tract (unpublished data). This was indirectly sup-
`ported in the current study by the fact that com-
`plete absorption relative to an ABT-089 solution
`was achieved with the hydrophilic matrix (HHH)
`and that absorption of a compound having a very
`short half-life was extended beyond the time
`frame of small intestine transit in dogs.
`In the fasted state, estimates of bioavailability
`relative to oral solution were 103.6 + 6.2% and
`59.9 ± 14.0% for the hydrophilic (HHH) and the
`hydrophobic (W3) matrix, respectively. The hy-
`drophobic matrix provided a smaller fluctuation
`ratio (c,„/C24 h) of the plasma levels. The more
`prolonged absorption corresponding to the lower
`extent of absorption was not unexpected as for-
`mulations with slower release rates could be ex-
`pelled from the body prior to the completion of
`the drug release. It is known that the average
`transit time in the entire gastrointestinal tract in
`fasted dogs is about 13 h (Davies and Morris,
`1993).
`The in vivo release of the drug from matrices
`was estimated by deconvolution of plasma profiles
`based on linear system analysis (Cutler, 1978).
`Plasma concentration data for each dog following
`
`administration of the oral solution were fitted to
`polyexponentials and used as unit impulse re-
`sponse C,5(t). Drug plasma data of the tablet
`formulations from the same dog, C(t), were de-
`convoluted with Co(t) to obtain in vivo drug
`release from the matrix formulations. The cumu-
`lative percent of the dose released in the GI tract,
`i.e. 'GI bioavailability' (Gillespie and Veng-
`Pedersen, 1985) of both matrices were estimated
`by the plateau values from respective in vivo
`release profiles at 24 h. The bioavailability esti-
`mates were 105.7 ± 6.5% for the hydrophilic sys-
`tem and 61.7 ± 13.8% for the hydrophobic
`system, respectively. The results closely match the
`relative bioavailability data calculated based on
`AUC, suggesting formulation release-controlled
`apparent absorption.
`
`3.3. In vitro /in vivo correlation
`
`The establishment of a relationship between the
`in vivo input rate and in vitro drug release from
`the dosage form is an important part in the
`formulation development process. A validated re-
`lationship between in vitro and in vivo release can
`be useful in setting up a meaningful in vitro
`quality control procedure that is predictive of
`product performance and can justify a change in
`formulation (Skelly et al., 1990).
`The preliminary in vitro/in vivo correlations for
`ABT-089 sustained-release tablets were explored
`by comparing the in vitro release data with in
`vivo drug release obtained from deconvolution.
`The results in Fig. 6 demonstrate linear correla-
`tions between in vivo and in vitro release for both
`matrix systems. A slope of 1.09 with an intercept
`of zero was obtained for the hydrophilic matrix,
`indicating a near 1:1 relationship. A slope of 0.89
`along with a near zero intercept also shows a
`good linear relationship between in vitro and in
`vivo release for the hydrophobic matrix. Overall,
`the in vivo drug release rate was slightly slower
`compared with the in vitro drug release. It has
`been well recognized that separate in vitro/in vivo
`correlations are generally obtained for different
`types of products of the same drug (Skelly et al.,
`1990).
`
`MYLAN Ex 1045, Page 9
`
`

`

`52 (cid:9)
`
`Y. Qiu et al. /International Journal of Pharmaceutics 157 (1997) 43-52
`
`4. Conclusions
`
`In summary, the present study demonstrated
`that prolonged oral delivery of ABT-089 with
`reduced fluctuation of plasma levels can be
`achieved using matrix systems. However, it is
`difficult to obtain sustained delivery over a longer
`period of time with hydrophilic matrix systems for
`a compound having extremely high solubility and
`diffusivity, such as ABT-089. Hydrophobic matrix
`is more effective in extending the release of this
`type of compound. Therefore, use of either insolu-
`ble heterogenous matrix or reservoir system may
`be necessary in order to develop an once-daily
`formulation of ABT-089 for humans.
`
`Acknowledgements
`
`Technical assistance from Lisa Ruiz is appreci-
`ated. The authors also would like to thank Drs
`P.K. Gupta and S. Silvestri for helpful discus-
`sions, Dr W. Gillespie for providing the PCD-
`CON program.
`
`References
`
`Arneric, S.P., Bannon, A.W., Briggs, C.A., Brioni, J.D., Buc-
`cafusco, J.J., Decker, M.W., Gopalakrishnan, M., Holla-
`day, M.W., Kyncl, J., Lin, N.H., Marsh, K.G., Qiu, Y.,
`Radek, R., Donnelly-Roberts, D.L., Sullivan, J.P.,
`Williams, M., 1996. ABT-089: an orally active cholinergic
`channel modulator (ChCM) with cognitive enhancing and
`neuroprotective activity. In: Becker, R., Giacobini, E.,
`Robert, P. (Eds.), Alzheimer's Disease: From Molecular
`Biology to Therapy. Birkhauser, Boston.
`Arneric, S.P., Campbell, J.E., Holladay, M.W., Johnson, P.,
`Lin, N.H., Marsh, K.G., Petersen, B., Qiu, Y., Roberts,
`E.M., Rodrigues, A.D., Sullivan, J.P., Trivedi, J.,
`Williams, M., submitted. ABT-089 [3-(2-(S)-pyrrolidinyl-
`methoxy)-2-methyl-pyridine]: III. An orally effective
`cholinergic channel modulator with potential once-a-day
`dosing and cardiovascular safety. Drug Development Re-
`search, in press.
`Cutler, D., 1978. Linear system analysis in pharmacokinetics.
`Journal of Pharmacokinetics and Biopharmaceutics 6,
`265 -282.
`Davies, B., Morris, T., 1993. Physiological parameters in
`laboratory animals and humans. Pharmaceutical Research
`10, 1093-1095.
`
`Desai, S.J., Simonelli, A.P., Higuchi, W.I., 1965. Investigation
`of factors influencing release of solid drug dispersed in
`inert matrices. Journal of Pharmaceutical Sciences 54,
`1459-1464.
`Foster, T.P., Parrott, EL., 1990. Release of highly water-solu-
`ble medicinal compounds from inert, heterogeneous ma-
`trixes. I: Physical mixture. Journal of Pharmaceutical
`Sciences 79, 806-810.
`Gillespie, W., Veng-Pedersen, P., 1985. A polyexponential
`deconvolution method. Evaluation of the 'gastrointestinal
`bioavailability' and mean in vivo dissolution time of some
`ibuprofen dosage forms. Journal of Pharmacokinetics and
`Biopharmaceutics 13, 289-307.
`Higuchi, T., 1963. Mechanism of sustained-action medication.
`Journal of Pharmaceutical Sciences 52, 1145-1150.
`Hildgen, P., McMullen, J.N., 1995. A new gradient matrix
`formulation and characterization. Journal of Controlled
`Release 34, 263-271.
`Hogan, J.E., 1989. Hydroxypropylmethylcellulose sustained
`release technology. Drug Development and Industrial
`Pharmacy 15, 975-999.
`Hui, Y., Marsh, K.C., 1995. Sensitive detection of selected
`cholinergic channel activators derived with 7-fluoro-4-ni-
`trobenzo-2-oxa-1,3-diazole. Journal of Pharmaceutical and
`Biomedical Analyses 14, 131-142.
`Lee, L., 1992. In: Kydonieus, A. (Ed.), Treatise on Controlled
`Drug Delivery: Diffusion-Controlled Matrix Systems.
`Marcel Dekker, New York, pp. 155-198.
`Lin, N.H., Gunn, D.E., Ryther, KB., Garvey, D.S., Donnelly-
`Roberts, D.L., Decker, M.W., Brioni, J.D., Buckley, M.J.,
`Rodrigues, A.D., Marsh, K.G., Anerson, D.J., Sullivan,
`J.P., Williams, M., Arneric, S.P., Holladay, M.W., 1997.
`Structure-activity studies on ABT-089: an orally bioavail-
`able 3-pyridyl ether cholinergic channel moderator with
`cognitive enhancing properties. Journal of Medicinal
`Chemistry 40, 385-390.
`Otsuka, M., Matsuda, Y., 1994. Controlled drug release of
`highly water-soluble pentoxifylline from time-limit disinte-
`gration-type wax matrix tablets. Pharmaceutical Research
`11, 351-354.
`Otsuka, M., Matsuda, Y., 1995. Programmable drug release of
`highly water-soluble pentoxifylline from dry-coated wax
`matrix tablets. Journal of Pharmaceutical Sciences 84,
`443-447.
`Scott, D.C., Hollenbeck, R.G., 1991. Design and manufacture
`of a zero-order sustained-release pellet dosage form
`through nonuniform drug distribution in a diffusional
`matrix. Pharmaceutical Research 8, 156-161.
`Skelly, J.P., et al., 1990. In vitro and in vivo testing and
`correlation for oral controlled/modified-release dosage
`forms. Pharmaceutical Research 7, 975-982.
`Williams, M., Arneric, S.P., 1996. Alzheimer's disease:
`Prospects for treatment in the next decades. In: Brioni,
`J.D., Decker, M.W. (Eds.), Alzheimer's Disease. Wiley-
`Liss, New York, 1996.
`
`MYLAN Ex 1045, Page 10
`
`